Final Report Summary - THINSI (Thin Si film based hybrid solar cells on low-cost substrates)
Executive Summary:
ThinSi is an ambitious project to develop a solar cell processing chain for high throughput, cost-effective manufacturing of thin film silicon based solar cells on low-cost silicon substrates. The Si powder based substrates are made on the basis of an innovative powder-to-substrate concept. In line with the current trends in PV, this project aims to reduce the cost of solar cells and modules compared to those made by the conventional wafer based approach. The solar cell structure, which was the main subject for all ThinSi solar cell fabrication related developments, is very similar to a conventional bulk crystalline Si solar cell. The "ThinSi" solar cell structures consists of a low-cost Si powder supporting substrate (processed within the "powder-to-substrate" approach) and an active high- quality layer on top, which serves as a solar cell base. Thus, thin Si film based, but at the same time "all silicon" structure substitutes the Si wafer. The final structure can be called also as a "Si wafer equivalent".
It has been demonstrated that advanced silicon based substrates can be processed from Si powder using following state-of-the-art innovative technologies: (i) Thermal spraying; (ii) Hot pressing; (iii) Spark plasma sintering (SPS); (iv) block casting growth using low-cost Si powder feedstock followed by sawing of the grown Si ingots. It was established that all Si substrates, processed in frame of the mentioned above innovative technologies, have good enough mechanical and electrical properties to be considered as a supporting part of the Si wafer equivalent structure.
In addition to well established chemical vapour deposition (CVD) process for deposition of thin (~20-30 µm) Si layers, a set of innovative lower cost processes have been developed to realise the new concept: (i) thermal spraying, (ii) magnetron sputtering and plasma enhanced chemical vapour deposition (PECVD). New tools for deposition of thin Si layers (active layers in "Si wafer equivalent" structure) have been developed. High temperature sputtering and PECVD tools have been developed for in-situ deposition of poly or crystalline Si layers. All tools are designed for large scale silicon deposition, with capability to process 156mm × 156mm Si wafers. Moreover, an advanced Electrostatic Spray Assisted Vapour Deposition (ESAVD) tool has been developed and offer a non-vacuum and cost-effective coating method for deposition of transparent conducting oxide films. AFM/Raman/Ellipsometer/SNOM combined measurement tool and relevant methodology for the analysis of solar cell structures on nano-scale have been developed.
Four approaches have been investigated to make thin Si layer based solar cells: (i) the ‘high-temperature approach’ based on implementation of diffused or epi- emitters, (ii) the ‘low-temperature approach’ (hetero-emitter), (iii) the ‘diffused emitter" combined with TCO, (iv) the exfoliation and bonding of epitaxial silicon foils on low-cost Si powder based substrates. The highest solar cell efficiency achieved on Si powder based substrates in frame of "high-temperature approach was 11.9%. By using the exfoliation and bonding of epitaxial silicon foils with an implemented Bragg reflector, free standing structures were obtained which could be processed to solar cells and bonded to low-cost Si powder based substrates with efficiencies over 14%. It is concluded that approach bases on exfoliation and bonding of epitaxial silicon foils is a promising advanced technology for prospective thin Si based solar cell processing on low-cost Si based substrates.
Project Context and Objectives:
4.1.2 Project context and objectives
Today, in many regions of the world, PV systems can produce electricity at less than twice the cost of electricity purchased from the grid. Indeed, parity has already been reached in Japan and is foreseen to be reached in 5-10 years in other regions all over the world.
The PV module market is dominated by crystalline silicon solar cells. It is important to note that about 40% of the silicon module cost is from the silicon wafers. Currently, the silicon feedstock demand for PV applications is enormous and with a growing PV industry the need for crystalline silicon will increase dramatically. Therefore, a major part of current research activity is concentrated on a search for alternative silicon based solar cell concepts with reduced consumption of high-purity silicon. In conventional Si wafer based solar cell technology, most of the Si material acts as a mechanical carrier for the solar cell structures. However, since most of the optical absorption in Si takes place in the upper 15-30 µm, it is sufficient to use only thin Si layers with thicknesses in this range. Indeed, if special optical confinement schemes are implemented, even thinner layers can be used. In general, thin-film photovoltaics are assumed to become a market dominating technology in the long term and any development in this field is extremely important for the EU industry.
ThinSi is an ambitious project to develop a solar cell processing chain for high throughput, cost-effective manufacturing of thin film silicon based solar cells on low-cost silicon substrates. The Si powder based substrates are made on the basis of an innovative powder-to-substrate concept. In line with the current trends in PV, this project aims to reduce the cost of solar cells and modules compared to those made by the conventional wafer based approach.
The solar cell structure, which was the main subject for all ThinSi solar cell fabrication related developments, is very similar to a conventional bulk crystalline Si solar cell. The "ThinSi" solar cell structures consists of a low-cost Si powder supporting substrate (processed within the "powder-to-substrate" approach) and an active high- quality layer on top, which serves as a solar cell base. Thus, thin Si film based, but at the same time "all silicon" structure substitutes the Si wafer. The final structure can be called also as a "Si wafer equivalent".
In addition to well established chemical vapour deposition (CVD) process for deposition of thin (~20-30 µm) Si layers, it was necessary to develop innovative lower cost processes in order to realise the new concept. The following state-of-the-art innovative technologies for processing of advanced silicon powder based substrates have been considered in frame of the ThinSi project using: (i) Thermal spraying; (ii) Hot pressing; (iii) Spark plasma sintering (SPS); (iv) Block casting growth using low-cost Si powder feedstock followed by sawing of the grown Si ingots. It was established that all Si substrates, processed in frame of the mentioned above innovative technologies, have good enough mechanical and electrical properties to be considered as a supporting part of the Si wafer equivalent structure.
New tools for the cost-effective deposition of thin Si layers (active layers in "Si wafer equivalent" structure) were under consideration in frame of the ThinSi project activity. In particular, high temperature sputtering and plasma enhanced chemical vapour deposition (PECVD) tools have been developed for in-situ deposition of poly or crystalline Si layers. All tools have been designed for large scale silicon deposition, with capability to process 156mm × 156mm Si wafers. Development of an advanced Electrostatic Spray Assisted Vapour Deposition (ESAVD) tool, which can offer a non-vacuum and cost-effective coating method for deposition of transparent conducting oxide films, has been considered as an important activity towards non-vacuum technology for the cost effective processing of Si based solar cells. AFM/Raman/Ellipsometer/SNOM combined measurement tool and relevant methodology for the analysis of solar cell structures on nano-scale were considered as an important input into ThinSi activity, since in many cases such analysis is required to investigate peculiarities of advanced solar cell structures.
Important to mention that all project developments, concerning innovative technologies and equipment prototypes were based on a requirement to scale up and transfer to production lines by the end of the project or in a realistic time in a future. New market opportunities for the SME and industrial partners were considered as an important issue, for both, production tool suppliers and for end-users of the technology.
One of the targets of the ThinSi activity was to reduce the thermal budget for deposition of high quality poly-Si thin layers, since the highly doped Si based substrates contain high concentrations of impurities. Normal deposition is usually done by epitaxial high temperature growth at 1000–1200ºC and the risk to contaminate the growing/ crystallized high quality Si active layer on such a substrate is rather high. Therefore, several approaches to address this issue have been considered: (i) Lower temperatures for the deposition of thin Si layers; (ii) Alternative to conventional CVD deposition methods for Si layers based on thermal spray, PECVD and magnetron sputtering; (iii) Gettering.
The ThinSi solar cell structure, being wafer based and based on silicon, is very similar to traditional cells. This will ensure a low acceptance threshold in the solar cell industry, which presently is 95 % based on crystalline Si4. In spite of these similarities, several essential bottlenecks and developments in low-cost process methods, and therefore in the whole process flow for the processing of poly-Si on low-quality Si substrates solar cell, must be addressed:
(i) Reduction of cost and simplification of processing of low-cost Si based substrates using a “powder-to-substrate” concept (bottleneck 1)
(ii) Methodology for cost-effective deposition of high quality active layers on low cost substrates, including Si base, emitter and TCO antireflection coating retaining a high cell efficiency (bottleneck 2)
(iii) Thorough understanding of the electronic properties of the deposited solar cell thin films and their interfaces as a function of deposition parameters (bottleneck 3)
(iv) Development of advanced methods for optical confinement (bottleneck 4)
(v) Implementation of “powder-to-substrate” concept based processes flow for the solar cell fabrication to existing industrial lines currently operating in frame of single Si wafer based approach (bottleneck 5).
The complexity of the solar cell device structure requires high-level expertise in modelling, design, materials technology and device processing. The project partners have been chosen so that all the required experience will be available to the project: silicon feedstock processing, ceramics technologies, deposition and crystallization of thin Si layers, deposition of TCO layers, and expertise on materials and interfaces.
ThinSi project has the following specific objectives:
Objective 1
• To develop innovative methods for fabrication of low-cost Si based substrates (elimination of bottleneck 1). The development contains:
o Fabrication of Si powder with the desired purity, crystallinity, conductivity (~0.01 Ωcm) and particle size (up to ~100 µm).
o Processing of low-cost Si substrates from metallurgical grade (mg) or metallurgical up-grade (umg) Si powder by ceramics technologies (tape casting and thermal spraying) with crystalline Si seeding grains with diameter ~100 µm.
o Post processing of Si powder based substrates prior to the formation of the thin Si solar cell bases.
o Manufacturing procedures for pilot scale processing of Si powder based substrates
Objective 2
• To develop a process flow and advanced equipment for cost-effective fabrication of thin Si solar cells on low-cost Si based substrates, and assembly into modules (elimination of bottleneck 2):
o Development and optimization of prototypes of high throughput advanced equipment for the cost effective processing of Si powder based supporting substrates and thin Si based solar cells on their basis
o Development and selection of cost effective processes, for the formation of the Si base, the emitter and the TCO based antireflection coatings (ARC), and methods for their stabilization
o Development of metallization schemes for large scale thin Si based solar cells
o Development of a process flow and cost effective processing of thin Si based solar cells with conversion efficiencies up to 15%.
o Assembly of the advanced solar cells into complete modules at a processing cost targeting ~1 Euro/Wp.
Objective 3
• To analyse the electronic properties of individual solar cell materials and their interfaces as well as the relationship between the deposition parameters and the device properties (elimination of bottleneck 3):
o Analysis of optical and electrical properties of individual solar cell structure layers
o Analysis of the relationship between the properties of individual layers and interfaces, and the final properties of solar cells
o Ab initio simulation of electronic structure of solar cell materials and interfaces in solar cell structures under investigation for some selected cases
o Development of the quality inspection specific equipment and methodology for the express analysis of solar cell structures on nano-scale
Objective 4
• To develop new methods for optical confinement (elimination of bottleneck 4):
o Development of an internal reflector in thin solar cells on low-cost substrates
o Development of advanced wet chemical and reactive ion etching (RIE) texturing methods for thin Si layers and low-cost substrates
o Application of plasmonic nano-particles enhanced light absorption in thin Si based solar cells
Objective 5
• To implement the “powder-to-substrate” concept based process flows to existing industrial lines (elimination of bottleneck 5).
o Development of innovative up-scaling technologies and equipment prototypes that will be transferred to pilot production lines by the end of the project targeting processing cost ~1 Euro/Wp.
o To create new market opportunities for the SME and industrial partners of the project
Project Results:
4.1.3 Main S&T results/foregrounds
WP1: Si powder synthesis and low-cost substrate processing
Since all ThinSi developments related to fabrication of innovative low-cost supporting substrates are based on implementation of Si powder, in WP1, Si powders with the appropriate purity and diameter have been synthesised and used for processing of the low-cost Si based supporting substrates. During WP1, a set of cost-effective processing technologies for fabrication of Si based supporting substrates has been developed. The developed processes can be considered as a background for an innovative “wafer equivalent” technology based on powder-to substrate concept.
The Si powder based substrates are made on the basis of an innovative powder-to-substrate concept. In line with the current trends in PV, this project aims to reduce the cost of solar cells and modules compared to those made by the conventional wafer based approach. The solar cell structure, which was the main subject for all ThinSi solar cell fabrication related developments, is very similar to a conventional bulk crystalline Si solar cell. The "ThinSi" solar cell structures consists of a low-cost Si powder supporting substrate (processed within the "powder-to-substrate" approach) and an active high- quality layer on top, which serves as a solar cell base. Thus, thin Si film based, but at the same time "all silicon" structure substitutes the Si wafer. The final structure can be called also as a "Si wafer equivalent".
In WP1 It has been demonstrated that advanced silicon based substrates can be processed from Si powder using following state-of-the-art innovative technologies: (i) Thermal spraying; (ii) Hot pressing; (iii) Spark plasma sintering (SPS); (iv) block casting growth using low-cost Si powder feedstock followed by sawing of the grown Si ingots. It was established that all Si substrates, processed in frame of the mentioned above innovative technologies, have good enough mechanical and electrical properties to be considered as a supporting part of the Si wafer equivalent structure.
Information from WP1 is used as a background for further studies in WP2- WP5.
Fabrication of Si powder with the desired purity, crystallinity and particle size
The main aim of this part of the project (WP1 related activity) was to identify and to develop processes, which can be used as a background for an innovative “wafer equivalent” technology based on powder-to substrate concept. Therefore essential part of the activity was concentrated on Fabrication of Si powder with the desired purity, crystallinity and particle size. Experience from the partners using the Si powder for thermal spraying, hot pressing and tape casting followed by a spark plasma sintering, showed that a special refinement and adjustment of the Si powder quality was needed for each Si wafer sintering method, both with regard to conductivity and particle size distribution. ELKEM Solar has therefore started work on optimising the powder production. ELKEM Solar has produced and delivered 9 different Si powder qualities (with particle diameters from ~10 to ~100 µm, un-doped and doped) for use in low cost substrates based on feedback from the consortium. Several routes were investigated to sintering tape casting silicon tapes at temperatures below the melting point of Si. It was established that a spark plasma sintering (SPS) as well as hot pressing of silicon powder with and without doping elements are promising methods to produce Si wafers with the desirable properties (dense structure and conductivity < 0.01 Ωcm). By use of this method, substrates which satisfied the requirements of the project have been produced. For the substrates both Al and B doped Si powder has been used. B and Al doping has been done by mechanically mixing the Si powder with B/Al at SINT, while B-doped Si powder also has been produced at ELKS.
ELKS has produced and delivered different Si powder qualities for use in low cost substrates based on feedback from the consortium. The powders are in general tailored based on the following learning points from consortium feedback:
- In terms of Fe and other metal elements in the silicon powder, ELKSM was looking for qualities between quality A (metallurgical grade Si) and ESS (Elkem Solar Silicon™, solar grade quality).
- Highly conductive B-doped Si powder, which is suitable for sintering of Si based substrates with resistivity of about 0.01 Ohm.cm and below.
- Thermal spraying techniques at Pyrogensis require PSD between 10 and 90 µm. To be sure that the powder does not hinder flow or block the feed to spraying equipment, ELKS has produced a 45 – 80 µm PSD range
- Hot pressing and SPS techniques require PSD < 120 µm. After having taken out 45 – 80 µm, 0 – 45 µm and 80 – 120 µm types of Si powders were fabricated at ELKS and delivered to SINTEF
Moreover, it has been established that for the spray pyrolysis based thin film production a PSD range of 20-30 µm was preferred. Some problems and difficulties occurred upon producing Si powder with such a tight PSD range in this size region. However, ELKS has managed to produce a 1 – 50 µm powder first and will has developed further separation methods for taking out the 20 – 30 µm particle size range from the initial Si powder mix. The production of such small particle sizes with this narrow size range is a true challenge. Finally, it was possible to produce 10 – 50 µm fines using special crushing and sieving methods.
Figure shows a sketch of the optimised production method for feed powder to making thin – film substrates using thermal spray or SPS methods:
Figure 3: Production method from bi-product to substrate powder
The method is optimized in terms of resistivity, particle size, production yield and chemical purity.
As an example Table 1 show powders with low impurity level produced at Elkem Solar. All samples are upgraded with a HF leaching process to remove impurities / contamination from crushing / grinding.
Table 1. Undoped upgraded Si powders.
Sample Treatment Fe
ppmw Al
ppmw Ca
ppmw Cr
ppmw Cu
ppmw Mn
ppmw Ti
ppmw Ni
ppmw P
ppmw
ENV450 Upgr. A+ 12.0 6.1 95 1.5 5.3 2.20 0.59 1.3 1.3
ENV451 Upgr. A+ 9.6 2.9 150 < 0.5 1.4 2.80 0.55 0.55 1.5
ENV452 Upgr. A+ 11.5 2.8 150 1.1 1.4 0.41 0.80 1.1 1.2
ENV448 Upgr. A+ 7.9 3.3 120 <0.5 1.8 0.38 0.50 NA 1.2
ENV456 Upgr. A 51 7.6 160 <0.5 5.1 0.92 0.86 0.69 1.5
ENV482 Upgr. A++ 8.6 2.6 75 5.5 13 1.4 0.27 1.6 1.6
From Table 1 it can be seen that the total impurity concentration somewhere between 150 and 170 ppmw (excluding oxygen).
Crystallinity of Si powder has been measured by micro-Raman. Results for Si powders A+75, A+40, A+10, D size1 and D size 2 can be seen in Figure 4.
Figure 4: micro-Raman spectra for different types of Si powders produced by Elkem Solar
It can be concluded that at least sub-surface region (~0.5 µm) of Si powder is not crystalline, although XRD measurements show that the main part of the powder is fully crystalline. This result shows that amorphous sub-surface powder regions can affect results of sintering and can cause problems to achieve highly doped Si substrates. Therefore such regions have to be re-crystallized upon the sintering.
Optimisation of processing conditions for fabrication of Si based supporting substrates by spark plasma sintering and hot pressing
After and in parallel with the successful developments related to fabrication of Si powders with the desirable properties, enormous efforts in ThinSi project were focused on processing of low cost Si powder based substrates with resistivity ~0.01 ohm•cm. It was established that the substrate quality obtained by tape casting followed by a solid phase sintering below the melting point of silicon was inadequate, too dusty and with too high resistance. It was concluded that stable substrates with satisfactory quality were very difficult to produce due to the difficulties with sintering of silicon. Therefore several alternative routes were investigated to improve the sintering of silicon tapes produced by tape casting. It was established that spark plasma sintering or hot pressing of silicon powder with and without doping elements are the most suitable methods to obtain dense and conductive Si powder based substrates. Moreover, conditions for the processing of Si powder based substrates by thermal spray method have been found.
Images of SPS based Si wafers are shown in Figure 5:
Figure 5: Spark plasma sintered silicon wafers (a) disc with d = 80 mm (b) disc with d = 200 mm.
It was established that to sinter good quality Si powder substrate a partial melting at the powder grains is required. This requirement causes some technical problems (melting of the central part of sintered wafer), which have been solved by optimisation of the SPS process. SEM image of the SPS Si powder based substrate is shown in Figure 6.
Figure 6: SEM images of two different locations on SPS Si powder sintered wafer surface
From Figure 6 it can be concluded that SPS Si powder based substrates have multi-Si like structure. Thus, Si powder based substrates processed by SPS method have structural properties, which are comparable with those for multi-Si material. Since the grain size in SPS material is much larger than the initial size of Si powder (< 100 µm), it can be concluded that partial melting of Si powder under SPS sintering conditions occurs.
It has been established also that in case of a low-cost Si powder (mg-Si) the SPS samples surface has depressions in which a very porous structure can be seen (Figure 7).
Figure 7: Depressions on the surface of SPS Si powder based sample
The depressions often contain distinct hexagonally shaped platelets. The particles are shown to contain chromium and are therefore probably CrSi2 which belong to the hexagonal crystal system.
Following this strategy (melting of Si powder), an ingot was made by melting and recrystallization of silicon powder (A+) mixed with boron providing the required resistivity of 0.01 ohm•cm. The experiment was done at SINT using a Crystalox DS-250 furnace. The material was melted in the furnace and slowly solidified with accurate control of the temperature profile and gas flow to guarantee ingot quality. The wafer cutting was done at Fraunhofer ISE using standard sawing techniques. 4-probe conductivity measurements were done giving an approximate resistivity of 0.006 ohm•cm. The wafers from this ingot were intensively used in WP4 for the solar cell processing.
In addition to mentioned above approaches, which includes melting of Si powder, hot pressing method, which was used at temperatures below melting point of Si has been tested as well. Si powder based substrates processed by hot pressing at high pressure and high temperature consist of a mix of small and very small crystallites (Figure 8).
Figure 8: Secondary electron image and orientation image micrograph (EBSD) for sample sintered by hot pressing.
From Figure 8 it can be seen that no twins or other Coincidence Site Lattice (CSL) boundaries observed, i.e. no recrystallization. No texture observed. Such hot pressed samples can be subjected to any wet chemical treatment prior any vacuum based process, since voids can't be seen on the surface of such sample.
Nevertheless, since sintering process can be done at different conditions, in some cases, if pressure and temperature are not optimized, hot pressed samples can contain some porous areas and application of wet chemical treatments for such samples is rather questionable.
It can be concluded that etching of hot pressed Si powder samples has some peculiarities, which depend on the sintering conditions. In general, wet chemical based treatments can be applied to such kind of Si powder based substrates.
Optimisation of processing conditions for fabrication of Si based supporting substrates made by thermal spray of Si powder
Several families of Si substrates made of low cost Si-based powders were produced. Some of them showed properties which fulfilled the selection criteria applied in the project: (i) low resistivity, below 0.01 ohm•cm, and (ii) acceptable consolidation and mechanical stability.
Production of free standing low resistivity (~0.01 ohm•cm) Si layers using thermal spray method has been performed using silicon powder size of 45-75 micron Quality D. The low resistivity was achieved by mixing in house, Si powder with Al and Boron dopants from 1% to 15%. The thermal spray technology used was Flame spray for higher production rates and cost effectiveness.
Free standing Si substrates with various dimensions from 50x65 mm2 to 156x156 mm2 and thicknesses of 300-700 µm have been processed (Figure 9).
The problems with the low mechanical stability and dusting from the substrates were overcome by applying a hotter thermal spray parameters and a post spray heat treatment protocol.
Figure 9: Si free standing low cost substrates with dimensions 50 x 65 and 156x156 mm2 processed by thermal spray
Important to note that Si substrates sintered by thermal spray after optimization of sintering conditions have crystalline structure, as measured by Raman (Figure 10).
Figure 10: Raman spectra of Si powder based wafer processed by thermal spray
Important to mention that the surface of Si powder based substrates sintered by thermal spray is rather rough (Figure 11a), which has some advantages for applications in solar cells, since such roughness provide dark colour and therefore good light trapping properties for such substrates.
Figure 11: (a) - Surface, and (b) - cross-section SEM images of the thermal spray sintered Si powder based wafer
At the same time, cross section SEM image shows that substrates include a lot of voids
(Figure 11b), which do not affect essentially mechanical properties of such substrates, since it seems that voids can block evolution of cracks in such structures.
Development of wet chemical methods for texturing and pre-treatment of low-cost Si based substrates prior to formation of high quality thin Si layers
Several types of SPS Si powder samples have been tested after CP4 (HF:HNO3) etch. Results can be seen in Figure 12.
Figure 12: Surface and cross section SEM images of SPS Si powder sample after CP4 (HF:HNO3) etch
From Figure 12 it can be seen that dense sintered Si regions and inclusions of not fully sintered agglomerated Si (potentially Si with contaminants, or dopants which is not fully reacted). Also cubical-like structures can be seen (right image) due to not fully sintered Si powder. As a result, formation of porous Si structures on such material could be difficult to realise.
Incorporation of porous Si layers into the Si supporting substrate
Porous Si etching in thick, 50 x50 mm2 multi-crystalline p+ substrates, made out of the A1 part of the ingot made of re-melted Si A+ powder (made by Sintef-ISE). The Substrates were first chemically polished to obtain a reasonably flat surface. A porous Si stack in the top part of the substrate was etched. Total reflectance measurements shows a clear reflection band around 800-900nm (theoretical reflection maximum at 950nm), which is a very good result for a first attempt of etching a porous Si Bragg reflector in an ‘unknown’ substrate. X-SEM image of the stack of porous Si layers is shown in Figure 13. This stack is creating an optical Bragg mirror etched structure.
Figure 13: X-SEM at top surface of SPS substrate after porous Si etching
From Figure 13 it can be seen that etching of a porous Si Bragg reflector is possible for the case of Si powder substrate subjected to a melting, which provides formation of a dense crystalline structure.
Results of the WP1 related activity can be summarised as follows:
• Optimisation of silicon powder production, including B-doping process to obtain highly doped Si powder based substrates
• Identification and demonstration of SPS and hot pressing methods for fabrication of dense and highly conductive silicon powder based substrates
• Identification of thermal spraying of silicon as a method to produce dense and highly conductive silicon substrates and thin layers
• Low cost wafers have been made from re-crystallisation of low purity silicon powder doped with Boron.
• Bragg reflectors based on porous Si structures have been etched into multi-Si wafers made from the ingot grown from low purity silicon powder doped with B.
WP2: Thin film processes
Si deposition by alternative methods
Epitaxial layers have been deposited at temperatures of 1050°C and 1000°C in the Atmospheric Pressure Chemical Vapour Deposition system RTCVD160 at the Fraunhofer ISE. Epitaxial layers with defect densities of < 105 cm-2 have been achieved at a temperature of 1050°C (Figure 14).
Figure 14: Sample surface after epitaxial deposition and treatment by “secco” etching. Defects like etch pits and stacking faults can be seen.
Both etch pit densities and stacking fault densities have been taken into account. Defect densities below 2.6•105 cm-2 were measured for epitaxial Si layers deposited at 1000°C.
OIPT has developed special tools, based on PECVD and magnetron sputtering processes, which allows deposition of Si based layers at temperatures up to 700 °C (PECVD) and 700 °C by magnetron sputtering. It has been demonstrated that deposition of Si layers on Si substrates at elevated temperatures (~700 °C) leads to a partial crystallization of such layers.
10-20 µm thick silicon films were deposited by pulsed DC magnetron sputtering with substrate table temperatures up to 690 °C on highly doped crystalline silicon and thermal oxide substrates. Deposition rates of silicon were measured by ellipsometry and lift-off and surface profile measurements. The a-Si samples were annealed at different table temperatures and for various times lengths. Raman spectroscopy and ellipsometry was utilized to confirm crystallinity of the silicon films. To demonstrate the possibility to use magnetron sputtering for the cost effective deposition of silicon layers on different substrates, two magnetron sputtering systems at Oxford Instruments have been used, the PlasmaPro System 100 and System 400.
The PlasmaPro System 100 is a single target system which can be configured with one pulsed DC, DC or RF magnetron up to 254 mm diameter. It is capable of loading circular wafers up to 200 mm diameter or 156 mm square. The target to substrate separation can be automatically varied from 11 cm to 19 cm while maintaining high vacuum, 3 × 10-7 Torr. RF bias on the substrate table allows plasma-assisted deposition and the capability to run pre-cleaning for greater adhesion and control of film structure and stoichiometry. The substrate temperature control is provided by an embedded heater, with a temperature range of 20 °C to 700 °C. Figure 15 shows a general overview of the PlasmaPro System 100.
Figure 15 A schematic of the PlasmaPro System 100.
The PlasmaPro System 400 can be configured with up to four magnetron targets to a maximum diameter of 200 mm. The magnetrons can be energised individually as either pulsed DC, DC or RF modes for batch or single-wafer processing. The single process chamber is partitioned into four sections, isolating the sources from each other without the high cost of clustering several single process chambers. A rotating shutter eliminates cross-contamination and enables target cleaning and sputtering stability before the wafers are exposed.
The system is capable of loading multiple substrates up to 200 mm round or 156 mm square. The wafer table may be operated in either static or rotating mode i.e. either with the wafer to be deposited placed under the energised target or continuously rotated through the deposition flux from the energised target. Uniformity masks can be fitted to optimise uniformity in rotating mode, or removed for high-rate deposition in static mode. Substrate RF bias offers plasma pre-cleaning and plasma-assisted deposition. The substrate table can be heated up to 300 °C. Figure 16 shows a general overview of the PlasmaPro System 400.
Figure 16 A schematic of the PlasmaPro System 400.
A major advantage of this system is that multiple materials can be deposited in one chamber in a single process. Multi-layer stacks are automated by process recipes in the software. The PlasmaPro System 400 provides the ability to sputter metals, semiconductors, as well as non-conducting materials with film thicknesses from sub nanometer up to many micrometers. It is well suited to the deposition of solar cell materials for which the backside reflector layer, silicon base and emitter layer, passivation layer and transparent conducting oxide layer can be deposited in this one chamber and in a single sequence.
Both sputtering systems are fitted with a wafer handler, either in the form of a single wafer load lock or batch cassette loading, since this minimises particle contamination and preserves the vacuum conditions of the process chamber.
Deposition process conditions: Two targets were used for the work in this paper, a lightly and a highly boron doped p-type silicon target, with a resistivity of 1 Ωcm and 0.01 Ωcm, respectively. Although both targets are conductive, pulsed DC mode was used. Pulsed DC differs from full DC by the addition of a reverse voltage bias oscillation with variable reverse time and pulse frequency. This reverse pulse is routinely used with dielectric films as it eliminates target arcing by attracting electrons in the plasma towards the target surface to maintain charge neutrality at the end of each deposition pulse. Highly doped silicon can be sputtered with full DC; however, pulsed DC is commonly used for silicon deposition to suppress arcs.
The process recipe includes four steps prior to deposition, a pump down step to 3 × 10-7 Torr to reduce contamination from atmospheric gas admitted during wafer loading, a dwell step for high temperature processes to enable the wafer to attain temperature equilibrium, an in-situ substrate clean using the table RF and a target clean (System 400 only). The DC power was set up to 3000 W during silicon deposition and was pulsed at 100 kHz for all runs with a reverse current time of 4 µs. For the majority of silicon samples, argon was the only process gas used and the process pressure was kept at 2 mTorr. This results in deposition of hydrogen-free silicon. However, the option to add hydrogen or other gases, for example phosphine and diborane for in-situ doping is feasible.
Thin silicon layers have been deposited with various table temperatures up to 690 °C on crystalline n-type silicon <100>, thermal oxide, aluminium and glass substrates for comparison. The table temperature was held to within 10 °C and the process parameters varied to optimize the silicon crystallinity. For a-Si deposition the deposition rate was optimized as a function of process parameters and target to substrate separation.
After deposition of silicon samples with various thicknesses, post-deposition annealing in vacuum at table temperatures up to 1000 °C for up to 3 hours was performed. The deposition rate was determined by lift-off and a surface step profile measurement (Ambios XP-300). Optical thickness and refractive index measurements were modelled using an ellipsometer (WVASE, J.A. Wollam). The fraction of crystalline structure was evaluated by means of micro-Raman spectroscopy measurements.
Results: ~1 µm thick silicon layers were deposited on crystalline n-type <100> silicon and thermal oxide with various table temperatures up to 690 °C. Figure 17 shows micro-Raman spectra for as-sputtered silicon at temperatures up to 690 °C. All the films exhibit crystallinity with small fractions of poly-crystalline silicon inclusions as shown by the shoulders in the range of 510-515 cm-1. It can be seen that silicon deposited on silicon substrates is more crystalline than that deposited onto thermal oxide.
Figure 17: Raman spectra for 1µm silicon deposited at 690 °C on thermal oxide (blue) and silicon (black) showing the crystalline phase.
Figure 18 shows Raman spectra of Si layers deposited on Si and glass substrates at 650 °C by PECVD.
Figure 18: Raman spectra of Si layers deposited by PECVD on Si and glass substrates at 650 °C.
From Figure 18 it can be seen that that almost complete crystallisation can be achieved upon in-situ PECVD deposition of Si layer on Si substrate at 650 °C.
Pyrogenesis has developed process for the deposition of thin (20-50 µm) layers by thermal spray using Solar grade Si powder. Cross section SEM image of such layer is shown in Figure 19.
Figure 19: Cross section SEM images of thermal sprayed Si /multi Si
From Figure 19 it can be seen that Si powder based layer deposietd on Si substrtes are rather dense and do not containe high concentraion of voids. Important to note also that such layers are fully crystalline ccording to Raman measuremnts.
Electrostatic Spray Assisted Vapour Deposition (ESAVD) of TCO layers
IMPT’s own the intellectual property of the eco-friendly and cost-effective non vacuum thin film deposition technique based on Electrostatic Spray Assisted Vapour Deposition (ESAVD). This method is an advanced version of the spray pyrolysis and a variant of the Chemical Vapour Deposition (CVD) process. ESAVD involves spraying atomised precursor droplets across an electric field where the charged droplets will undergo decomposition and chemical reactions in the vapour phase and deposit onto a heated substrate. The homogeneous and/or heterogeneous chemical reactions can be tailored to synthesise the desired films for various applications. ESAVD is a disruptive coating technology. A novel atmospheric deposition technique and tooling based on ESAVD has been developed for the deposition of transparent conducting oxides to be applied to thin film silicon solar cells. A low cost and non vacuum tool has been developed based on ESAVD for the production of transparent conducting oxides thin films. This new ESAVD tool provides a non vacuum and cost-effective coating method and process which might not be realised by the existing available tools.
A prototype scale-up ESAVD system (see Figure 20) has been researched and developed for this project as a deposition tool with the following features:
a) Non vacuum atmospheric system
b) Wafer load up to 200 mm round/square.
c) High temperature platen to accommodate up to 200 mm square Si wafer heating up to 600°C
d) Platen carrier equipment to integrate with R&D scale-up machinery
e) Motion stage equipment
f) Adjustable spray atomiser and wafer substrate working distance
g) Dedicated spray atomiser and assembly
h) Spray motion control software developed
i) High voltage device up to 20kV
j) Chemical liquid delivery system sufficient for dispensing precursors during large wafer deposition.
Figure 20: A prototype scale-up ESAVD system
The optical and electrical properties of the ITO film deposited by ESAVD are summarised as follows: sheet resistance – 47 Ω/•; resistivity – 4.3x10-4; carrier concentration 5.4x1020 e/cm3; mobility 27.3 cm2/Vs; optical transmission > 88%; refractive index – 1.8-2.0.
Figure 21: Hardware of the prototype scale-up ESAVD system for processing large area ITO onto Si wafers.
Results of WP2 related activity can be summarised as follows:
• OIPT developed advanced magnetron sputtering tools, which provide possibility to deposit crystalline or polycrystalline Si layers in an in-situ process at temperatures around 700°C.
• PyroGenesis has demonstrated production of thin Si coatings of 10-15 micron thickness using solar grade Si powder for the thermal spray deposition method.
• A prototype scale-up ESAVD system has been developed
• Highly conductive ITO layers as well as several alternative TCO films have been deposited by ESAVD and magnetron sputtering methods.
• OIPT has develop processes for deposition of thick Si layers (10µm), (poly)/crystalline emitter layers (100-300nm), and thin emitter layer (5-20nm) for integration into Si based solar cell structures.
WP3: Analysis and properties of thin layers and solar cell structures
Analysis of composition and electrical properties of individual Si based supporting
substrates and thin Si layers
Properties of Si supporting substrates and thin layers is systematically analysed by XPS, XRD, SEM, Raman, conductivity, SIMS.
Moreover, an integrated "AFM/Raman/Ellipsometer combined measurement tool" has been developed by NT-MDTin order to characterize Si based solar cell structures on nano-scale. The combined tool consists of two systems. The first is the integrated AFM/Raman system with the optional SNOM option, and another is the integrated AFM/Ellipsometer systems.
Integrated system is equipped with three switchable ‘laser’ inputs, which provide the possibilities of analysis on photo-voltaic specimen on a nanoscale with various wavelength lasers. This state-of-art system can be updated friendly with a new type of digital controller, which enables to perform ‘SKPM and SCM’ methods in-situ. These techniques provide the abilities of conductivity analysis on both ‘electric and dielectric’ surface solar cells and Photovoltaic in combined with Raman analysis.
AFM/Raman integrated system mainly consists of:
• Integral Base,
• 100x100x10um closed-loop scanner,
• Optical AFM measuring head with the inserted cantilever holders, which allows use the Hollow cantilever for SNOM operation,
• Raman Spectrum with three lasers options of 473nm, 532nm, and 633nm, and the other lasers of 405nm, 488nm, and 785nm can be easily installed as well,
• Laser Radiation input module,
• Detectors – CCD camera and PMT (updating standard CCD to EMCCD for faster Raman Mapping)
• Optical viewing system with 100x, 0.7 NA, objective
• SPM controllers (updating to the digital controller, which allows the SKPM/SCM measurement simultaneously for conducting and dielectric layers of solar cells)
This system provides the full AFM (Topography/morphology, conductivity, photo-current, potential) characterization on the solar cell (front surface and cross-section) together with the Raman characterization.
AFM/Ellipsometer integrated system
For integration of AFM with ellipsometry, a Woollam M-2000 spectroscopic ellipsometer was used by NT-MDT. The whole stage was designed in order to integrate with AFM system including to the ‘adapter stage and ARM support’, which enables ‘Full AFM measurements on the solar cell cross-section surface and Ellipsometer measurement from the aside under the AFM probe’ (Figure 22 a). The optical light indent/receipt tube with fine adjusted screws for optical alignment was modified and the final diameter of these two optical tubes was modified from 1mm down to 10um, shown in Figure 22b. Such modification allows us to analysis the multi-layers on nano-meter scale, especially for textural solar cell with eliminating the surface roughness effects.
Figure 22: a) top-view of the design for the integrated AFM/Ellipsometer system; and b) the indication of the modified optical indent tube
This final integrated system mainly consists of ‘AFM basement, Smena AFM head, AFM controllers and Ellipsometer equipped with separate controller and software’. Such integrated tool allows to complete AFM measurements and analysis on the cross-section surface of solar cells.
Analysis of properties of TCO and passivating layers
A special methodology for optical characterization of individual layers, interfaces and final solar cell structures, based on traditional spectrophotometry and spectral multi-angle ellipsometry, within which novel approaches both for spatially resolved measurement, and for modeling of solar cells multilayers is being developed. Passivation properties of thin layers will be analyzed by QssPC and μ-PCD methods for which the procedures have be already prepared.
Hydrogen related phenomena at ITO/a-Si:H/Si heterojunction solar cell interfaces have been studied as well as advanced ellipsometric characterization of single layers and solar cell structures was performed preliminarily for bulk and surface in homogeneity of ITO and AZO from ENEA.
Solar structures of the type ITO(~80nm)/p+-a:Si:H(~5nm)/polished n-Si (~250mm) have been analysed for their optical and geometrical parameters. The depth profile modelling of the optical constants has yielded important information on free-carriers distribution within the film thickness.
Ultrathin a-Si layers from OIPT: a vast set of PECVD-deposited a-Si layers analyzed provides a data base for the analysis of the correlation of the film properties with the deposition conditions.
Analysis of interface properties of solar cell structures
Analysis of interface properties of the solar cell structures has been performed using SEM and IV measurements. In some cases TO/Si interface has been studied by XPS and TEM. SSRM and SCM measurements have been used for the analytical investigations and control of solar cell structures on a nano-scale.
Interface properties of Solar cell structures processed using passivation by dielectic oxides
A set of solar cell structures processed on the bases of Si powder based substrates was investigated to clarify influence of interfaces on their properties.
Below a brief overview of interface properties of a typical solar cell stricture processed at ISE Freiburg is given.
Solar cell structure “Hc2c” : highly doped Cz substrate with ~ 2 µm p++ BSF and ~ 22 µm p-type base, no HCl-gettering, no texturing, diffused 80 Ohm/sq emitter without oxide passivation and directly evaporated Ti/Pd/Ag front metallization, double layer anitreflection (TiOx, MgFx), processed in the "ISE solar cell run", eta=10.6%, VOC=609mV, JSC=26.0mA/cm2 FF=67.4%
Interface of “Hc2c” solar cell structure is shown in Figure 23
Figure 23: TEM image of MgO/TiO/epi-Si/Si solar cell structure
From Figure 23 it can be concluded that at the interface between TiO passivating layer and Si a thin SiO2 layer is present, which is formed upon deposition of TiO oxide. It can be assumed that such layer can affect efficiency of solar cells, since passivation properties of TiO can be affected by such a layer. More detailed studies are required to clarify this interface related problem.
Interface properties of Solar cell structures containing TCO layers
Experimental: The Indium oxide (IO)-based films were deposited by the pyrosol method. The film-forming solution for ITO contained 0.2M InCl3 /3%SnCl4 /10M H2O in methanol, for IFO - 0.2M InCl3/0.15M NH4FHF/0.8M H2O in methanol. The ultrasonic nebulizer operating frequency was 2.64 MHz. The deposition temperature varied in the range of 390-490º C. Argon was used as a carrier gas.
Transmission electron microscopy (TEM) as well as energy filtered transmission electron microscopy (EFTEM), was performed using with a JEOL 2010F instrument, equipped with a GATAN energy filter on standard cross section specimens of the films.
N-type Cz Si wafers with a resistivity of 4.5 Ωcm and 390 µm in thickness were used for the SC processing. The wafers were textured on the front side. Fabrication of p+nn+ structure was performed by diffusion from deposited boron and phosphorous containing glasses at temperatures around 1000ºC. After the removal of glasses all samples were etched in a HF-HNO3 solution to make the emitter shallower. ITO thin (~80 nm) layers were deposited on the p+ and the IFO layers on n+ side. The ITO films had a sheet resistance of about 50 Ω/sq and the IFO films of about 30 Ω/sq. After TCO deposition, samples 6.5×6.5 cm2 in size were cut.
Figure 24 shows interface region between ITO and Si substrate
Figure 24: TEM images of ITO/Si interface for the case of ITO layer deposited at 400 ºC.
From Figure 24 it can be seen that a buffer layer between ITO and Si substrate is formed. The composition of this layer has to be investigated in details, which is not a subject of this work. Nevertheless it can be assumed that most probably this ~2 nm thick layer consists of silicon oxide, which is formed upon the ITO deposition at quite high temperatures in presence of water.
It can be concluded that at the interface between ITO as well as TiO layers and Si a thin SiO2 layer is present, which is formed upon deposition. Formation of such layers has to be taken into account for the analysis of the solar cells properties.
Analysis of solar cells and solar modules properties
Conversion efficiencies of solar cells were characterized and certificated by ISE. Spectral response, dark and illuminated IV measurements were already performed for solar cells processed within the consortium.
Results of WP3 related activity can be summarised as follows:
• Integrated prototype system AFM/RAMAN/Ellipsometry has been developed within WP3 by T-MDT. The tool provides the Raman, Topography, Photo-current, et al done on the same area simultaneously. Topography surface characterization in a larger Z scale and higher resolution as well all the detailed topography investigation, conductivity measurements, higher resolution (SSRM), all is possible with the tool. Currently demonstration measurements are in progress.
• Detailed optical and electrical analysis of the porous Si multilayer structure, made out of the re-melted Si A+ powder acting as reflector. A clear reflection band can be seen around 800-900 nm, by total reflection measurements. Very uniform multilayer structure also visible by cross section SEM.
• Detailed analysis (SEM) of a front surface texturization based on plasma self masking process. Detailed analysis of the surface and interface between active layer and reflector as well as top surface after texturization (SEM). Very good uniformity, low reflectivity at desired wavelengths and low Si removal down to 0.6um were measured repeatedly.
• Structures of the type ITO(~80nm)/p+-a:Si:H(~5nm)/polished n-Si (~250mm) have been analysed for their optical and geometrical parameters. The depth profile modelling of the optical constants has yielded important information on free-carrieris distribution within the film thickness.
• The SEM-EDX study of lowly and highly doped SPS substrates reveals that the sintered silicon contains structures with a different composition. The grains with weak EBSD quality rise difficulty in detecting the crystal structure. Those structures are crystalline but probably highly stressed. The orientation of the EBSD pattern remains the same within the same large structure. EDX reveals Si, B rich regions and Si, B, O rich regions. In addition, intra-grain defects/dislocations in the crystalline Si are revealed by the microscopy after secco etch, indication a poor crystal quality. In general it can be said that the lowly doped Si SPS substrate quality is low.
• SIMS analysis of the Si based supporting substrate, made of recrystallised A+ powder, have been performed for B, C and O levels presence in the substrate. It was found that the O level is around 2E18 at/cm3 or less, there seems to be a gradient of C present in the top 300nm near the surface; from then on the C concentration stabilises at around 9E17 at/cm3 (well above the C detection limit of 2E17 at/cm3). The high Boron concentration was confirmed (p+ substrate) at about 2E19 at/cm3.
• Reference Epi-wafer equivalent solar cell with efficiencies of 14.5% was reached on monocrystalline Si substrates and 13% efficiencies were reached for powder based substrates. FZ material processed in the same batch resulted in solar cell efficiencies of 17%.
WP4: Design, processing, characterization and testing of solar cells and modules
Different approaches have been investigated in the project to make efficient thin Si layer based solar cells. In the conventional high-temperature approach, processes for the deposition of thin silicon epitaxial layers were applied. The junction was formed either by conventional phosphorous diffusion or by epitaxial deposition of the emitter. The second approach was the low-temperature heterojunction approach. In this approach, an amorphous silicon or silicon carbide layer was used as emitter which allows processes at low temperatures (< 250°C) and provide an outstanding surface passivation. Higher open circuit voltages (VOC) than with conventional diffused emitters are possible. The so-called HIT structure (Heterojunction with Intrinsic Thin layer) was used to achieve the best performance. For the third approach, a phosphorous diffused emitter was combined with a transparent conducting oxide (TCO). Therefore, indium tin oxide (ITO) was deposited on top of the thin silicon layer based structure. The fourth approach was the exfoliation and bonding approach, which uses porous silicon layers to detach high-quality epitaxial silicon layers and bond them to low-cost silicon powder based Si substrates.
Process chains
High-temperature approach: In the high-temperature approach, the epitaxial deposition of a thin crystalline silicon layer is done at ISE in atmospheric pressure rapid thermal chemical vapor deposition reactors at temperatures between 1100°C and 1200°C. High-quality epitaxial layers can be processed. The emitter is formed by two different methods in the high-temperature approach, either by emitter epitaxy or by emitter diffusion. Several solar cell runs were performed using the high-temperature approach.
The solar cell processing is described by the following steps:
• HCl gas gettering on Si substrates
• Epitaxial deposition by APCVD (Atmospheric Pressure Chemical Vapour Deposition) of ~ 2.5 µm Back Surface Field (BSF) and ~ 30 µm base
• Diffused phosphorous emitter/epitaxy (sheet resistance: ~80 Ohm/sq)
• Al back contact (full area, ~ 3 µm thickness)
• Directly evaporated front grid (instead of a photolithographically defined grid)
• Light induced metal plating
• Double layer anti reflection coating (DL-ARC)
• Edge isolation
Prior to the epitaxy process step, the samples with porous Si shortly HF dipped (with HCl added), the others were cleaned by “IMEC clean”. By high-temperature epitaxy, a back surface field (BSF) with a thickness of ~2.5 µm and a epitaxial base with a thickness of ~ 30 µm was deposited, after a pre-annealing step in H2 to close and reorganize the pores.
The “First joint IMEC-ISE solar cell run” was done using low-cost crystalline Si substrates and reference substrates. The goal was to assess the gettering efficiency of porous Si layers and the gettering efficiency of HCl gas gettering and assess the upper level of the conversion efficiency of such structures. Solar cells on p+ 99.94% UMG-Si substrates from ISE were processed together with p+ multicrystalline and p+ monocrystalline Si reference substrates.
In the next run named “ISE solar cell run” first solar cells on Si powder based substrates were processed applying a simplified cell process. It means that basically the process chain used without the porous silicon formation step and the plasma texturing. Also, instead of photolithography, the front contacts were directly evaporated leading to higher shading due to higher metallization finger thickness. Wafers of the ThinSi ingot A1, which was crystallized from Elkem Solar Si A+ powder at SINTEF and cut into wafers at ISE, were used as Si powder based substrates.
In the next solar cell run (“Second joint IMEC-ISE solar cell run”), Si powder based substrates were used again. To prevent diffusion of detrimental metals from the Si powder based substrate into the epitaxially deposited active layer, the porous Si layers were included again in this run.
Table 2: Results of the first joint IMEC-ISE solar cell run
Sample No. Substrate material Cell size [mm2] VOC
[mV] JSC [mA/cm2] FF
[%] ηcell
[%]
172A UMG 616 28.0 77.5 13.4*
175A UMG 620 28.6 76.5 13.6*
S4a mc (reference) 48 x 48 611 29.0 74.8 13.3*
M3 mono (reference) 48 x 48 644 30.4 78.4 15.3*
* independently confirmed by ISE CalLab
Whisker growing was observed during the epitaxial deposition step (see Figure 25). It was due to the high metal content in the substrates and resulted in shunting of the p-n junction and therefore in poor cell results (conversion efficiencies between 0.4% and 4.3%).
Figure 25: Whiskers can be seen on the ingot A1 substrates after epitaxial deposition of Si (even with applied HCl gas gettering).
It can be concluded that the metal contents in the Si powder based substrates are too high, still after HCl gas gettering. Substrates with lower initial metal concentrations have to be used and additionally, a porous Si layer from IMEC has to be integrated which acts as a diffusion barrier and prevents diffusion of detrimental metals from the Si powder based substrate into the active layer. Porous Si layers were used in the Second joint IMEC-ISE solar cell run.
To prevent diffusion of detrimental metals from the Si powder based substrate into the epitaxially deposited active layer, the porous Si layers were included again in the “Second joint IMEC-ISE solar cell run”. An overview of the results is given in Table 3. The porous Si layer can be used to getter detrimental metals within the substrate. It was reported recently by IMEC that there is a gettering effect by porous silicon. The gettering effect of HCl gas gettering was also reported earlier, the latest results were published within the ThinSi project in the special issue “Advanced concepts for silicon PV”. HCl gas gettering as well as porous Si gettering were applied in this cell run.
Unfortunately, no statistically relevant results could be found in the groups of solar cells including HCl gas gettering and porous Si layer etching (groups B, E, and G). One reason for the bad statistics in the cell results is a high number of broken edges and samples and broken cells. This might be due to the inferior material quality. Nevertheless, it could be demonstrated that the texturing is beneficial to the short circuit current and therefore to the cell efficiency. The highest solar cell efficiency achieved on Si powder based substrates was 11.9%. All cells of group F showed efficiencies above 10%, which means that milestone M11 has been reached.
Table 3. Cell results of the Second joint IMEC-ISE solar cell run
Group Wafer ID HCl gettering Porous Si Reflector Texture JSC [mA/cm2] Cell Efficiency η
B 5-6-7-8 YES NO YES 22.8 – 27.5 0.0 – 9.4
E 17-18-19-20 NO YES NO 9.9 – 24.3 0.4 – 3.6
F 21-22-23-24 NO NO YES 29.3 – 29.6 11.0 – 11.9
G 25-26-27-28 NO YES YES 24.6 – 26.1 1.5 – 1.8
H 29-30-31-32 N NO NO 26.0 – 28.4 4.6 – 9.6
Low-temperature approach: In the low-temperature approach solar cell runs were performed using the heterojunction approach. The summarized results of the HIT and (for reference) diffused emitter cells are presented in Table 4. The average value of the short circuit current (JSC) is about 2 mA/cm2 higher for the cells with diffused emitter because of a more effective anti-reflection coating (SiNx instead of ITO used for the HIT cells). The current densities of both types of cells are typical for a base thickness of 25µm without texturing, coated with an ARC. The VOC values of the HIT cells show a clear gain of 25 mV compared to the VOCs of the 80 Ω/sq diffused emitter cells. An average FF of 80% is obtained for HIT cells. Because of a lithography step which is not optimized for non-textured samples, some adhesion problems occurred during the lift-off process of the metal after evaporation. Therefore, the fill factor (FF) values for the cells with standard diffused emitter are lower than would be expected. In previous experiments, no difference was observed in the FF values for samples with HIT structure and diffused emitter structure.
For thin Si layer based solar cells on reference monocrystalline Cz Si, it could be demonstrated that the a-Si:H heterojunction emitter with intrinsic thin layer approach is beneficial to the VOC resulting in higher cell efficiencies compared to cells with standard diffused emitter.
Table 4: Comparison of the HIT cell and diffused emitter cell results, processed at IMEC
No. JSC VOC FF Efficiency η
[mA/cm2] [mV] [%] [%]
HIT cells 26.8 655 80 14.1
Diffused emitter cells 28.9 630 67 12.3
Diffused emitter with TCO approach
There were also structures processed within the project which consisted of an epitaxial Si base layer on low-cost Si substrates with a conventional diffused emitter and a transparent conductive oxide (TCO) at the front, with the latter deposited at UNOTT & IMPT. Some of the samples were processed to solar cells which were characterized at IMEC. The following substrate materials were used for the sample structures: ingot A1 p+ (made from A+ powder), ISE UMG p+ (as low-cost reference), and Cz p+. The substrate size was 50 x 50 mm2 for each structure. It is demonstrated that the best solar cell efficiency could be as high as 9.7%, which is comparable with the conventional technology, currently available at ISE.
Exfoliation and bonding of epitaxial silicon foils approach
Another approach which was investigated in ThinSi is the technique to exfoliate a thin silicon layer and bond it to a low-cost substrate. A high-quality monocrystalline silicon substrate can be used to process a so-called ‘separation layer’ by electrochemical anodization. Porosification is performed in such a way that a low-porosity layer and a high-porosity layer are created. After annealing in H2, reorganization of the pores occurs. The back surface field (BSF), the emitter, and the base can be deposited in-situ. A complete structure is presented in Figure 26.
Figure 26: Substrate with separation layer, low-porosity layer, and epitaxial cell structure for exfoliation.
After the separation layer (weak layer) formation and the epitaxial growth processes, the front surface processing is done while the epitaxial structure is still attached to the substrate. Afterwards, the structure is detached and the rear-side processing is done on the free standing foil. All steps of the exfoliation and bonding approach are shown in Figure 27.
Figure 27: Exfoliation and bonding of epitaxial silicon foils on low-cost substrates
Additionally, in this approach light confinement can be applied in terms of a bragg reflector, which means a stack of porous silicon layers of alternating low and high porosity. This feature leads to light trapping and thus an increased short circuit current JSC.
All process steps concerning porous silicon layer formation, exfoliation and bonding processes were performed at IMEC. The first solar cell batch of epitaxial foils in ThinSi was processed without Bragg reflector. The thickness of the active base was ~ 28 µm. The results of the free standing epitaxial foils before and after bonding are presented in Table 5. The best cells show efficiencies close to 12%. For the bonding on the various substrates, Ag based glue was used. There were some problems observed: Shunts could be observed on the edges of the cells, which are due to the liquid Ag based glue. That is why the cells were diced out after bonding. Also, handling problems of the free standing foils made the processing difficult.
Table 5: Results of free standing and bonded foils without Bragg reflector epitaxial foils
Device JSC VOC FF Eta
[mA/cm2] [mV] [%] [%]
Free standing cells I_free 19.0 621 73 8.6
II_free 25.4 628 75 12.0
III_free 24.2 626 75 11.4
IV_free 23.8 627 77 11.5
Bonded cells
I_bond 18.6 610 72 8.1
II_bond 25.6 627 75 11.9
III_bond Broken
IV_bond 24.2 626 75 11.4
For the second batch ‘advanced epifoils’ were created including a Bragg reflector and a separation layer, both prepared by electrochemical anodization. The thickness of the active base was 35 µm. The cells were diced out for measuring to minimize losses caused by shunts.
In Table 6 the results of advanced epifoils are presented before and after bonding. After bonding an increase of the JSC was observed, which is due to light trapping. But at the same time a slight decrease of the VOC occurred. The best efficiency of a free standing epifoil was 14.5%, whereas the best efficiency after bonding was 14.2%. The best efficiency of an epifoil bonded on the sintered substrate fabricated during the ThinSi project is 14.0%. This is close to the value which is requested in milestone M8 (~15%).
Table 6: Results of advanced epitaxial foils with included Bragg reflector before and after bonding
JSC VOC FF Efficiency η
[mA/cm2] [mV] [%] [%]
Epifoil 31.5 615 75 14.5
Epifoil on low-cost multi-c substrate 31.6 614 73 14.2
Epifoil 30.0 618 76 14.1
Epifoil on sintered substrate 30.3 617 75 14.0
Finally, mini-module was fabricated from thin silicon based solar cells, processed in frame of the project. Low-cost 99.97% UMG silicon substrates with size 100x100 mm2 were used. The efficiency of the active area of the mini-module is ~7.9%, which is well below the requirements for Si wafer/wafer equivalent based PV. It can be concluded that the large scale epitaxial growth processes still have to be improved and optimized, prior the implementation of wafer equivalent technology to industrial lines. It has to be noted, however, that all laboratory scale developments demonstrate that the advanced technology under the study is promising and needs further efforts for the up-scaling.
WP5: Transfer to production lines
In this report we present a cost model developed to analysis the industrial feasibility of low-cost substrate technology. The standard technology cost model is shown because of it is used as reference. General hypotheses used to build the cost model are presented. The cost model results of standard BSF-Al technology are presented and the industrial feasibility of this technology is analyzed. The concepts “return of investment” and the “annual relative profits” are introduced and will be used to analyze the industrial feasibility of the technology.
The process flow for standard technology and low-cost substrate approach are shown. The cost model results for low-cost substrate technology are presented and the industrial feasibility analysis is described. Efficiency thresholds at mass production level to balance the operational cost, the return of investment and the annual profits of standard technology using low-cost substrate technology are found. The cost model results for low cost substrate technology are shown in the following figure. We present three different substrate costs (30%, 50% and 70% of standard Cz-Si, that is 1.22 €/piece) and the standard BSF-Al technology, as reference; we show the sensibility of operational cost at 50 MWp production line scale versus average solar cell efficiency (Figure 28).
Figure 28. Operational cost for PV module based on low cost substrate technology at 50 MWp production line scenario. Three different substrate costs have been analyzed (30%, 50% and 70% of Cz-Si wafer cost) and the sensibility versus solar cell efficiency is shown. The operation cost of standard BSF-Al technology is shown as reference (green).
When substrate cost is 50% of Cz-Si wafer, the solar cell efficiency should be higher than 16.5% to assure a competitive operational cost; and when substrate cost is 30% of Cz-Si wafer, the efficiency solar cell at 50 MWp production line scale should be higher than 15.3% to assure a competitive operational cost for technology.
An estimation of exploitation cost for 500 MWp – 1 GWp factory scale have been done (Figure 29). For the BSF-Al standard technology, the operational cost for PV module is 0.58 €/Wp, assuming 17.5% as average solar cell efficiency. To balance this operational cost and this solar cell efficiency level using LCS technology is needed a 62% substrate cost of standard Cz-Si wafer for 1 GWp scenario.
Figure 29. Operational cost for PV module based on low cost substrate technology at 500 MWp – 1 GWp factory scenario. Three different substrate costs have been analyzed (30%, 50% and 70% of Cz-Si wafer cost) and the sensibility versus solar cell efficiency is shown. The operation cost of standard BSF-Al technology is shown as reference (green).
When substrate cost is 50% of Cz-Si, the solar cell efficiency should be higher than 16.8% to assure a competitive operational cost; and when substrate cost is 30% of Cz-Si wafer, the efficiency solar cell at 1 GWp production line scale should be higher than 15.5% to assure a competitive operational cost for LCS technology.
To estimate and to verify compatibility of the developed innovative processes on an industrial level, some technological developments have been performed at ISO in collaboration with IPMT and OIPT. In particular, it was demonstrated that ITO as well as a-Si:H layers can be deposited by magnetron sputtering on 156x156 mm2 Si wafers provided by Isofoton with the sufficient for industrial processes homogeneity.
Some attempts have been done to process large scale solar cells with ITO antireflection coatings at Isofoton. It has been concluded that screen printing metallization currently available at Isofoton, has to be modified in order to be compatible with the TCO based antireflection coatings, which are present in the most versions of the advanced solar cell structures developed in frame of the ThinSi project. In particular, the firing temperature has to be reduced up to ~400 ⁰C from currently available ~800 ⁰C. Development of a special metallization process on industrial scale requires additional efforts, time and resources, which were limited and not sufficient in frame of ThinSi.
Since large scale epitaxial based solar cell structures processed at ISE did not show sufficiently good efficiencies, it was concluded that optimisation of the basic processes for epitaxial deposition of thin Si layers still has to be done prior the testing of such structures on a pilot scale/industrial level.
Potential Impact:
4.1.4 Potential impact
Strategic impact
The impact of ThinSi conforms to the very ambitious targets that have been adopted for renewable energy in Europe, for example:
• In 2001, EU aimed to have 21% of its electricity coming from renewable energy sources by 2010 .
• In March 2007, the Heads of States and Governments of the 27 EU Member States adopted a binding target of 20% renewable energy from final energy consumption by 2020 .
• In September 2008, the European Photovoltaic Industry Association declared that photovoltaic energy could provide 12% of European electricity by 20202.
Although reliable PV systems are already commercially available and widely deployed, the cost of PV generated electricity is still too high to compete with electricity from non-renewable sources. Therefore, further development of PV technology with the aim to drastically reduce the turn-key system prices and cost/m2 cell area is crucial. Indeed, it was emphasised already in 2005 by the Photovoltaic Technology Research Advisory Council (PV TRAC) that such a development is possible.
For conventional silicon PV technology experience has shown that the price is reduced by 20% each time there is a doubling of installed capacity. As a consequence of this rule, a huge and heavily subsidised capacity would have to be installed if the required cost reduction should be realised by the economy of scale alone. Therefore a step-change in the PV performance is required.
ThinSi project demonstrated a complete process chain for cost effective processing of solar cells based on a “powder-to-substrate” approach. Using a cost model, developed by Isofoton it is estimated that when substrate cost is 30% of Cz-Si wafer, the efficiency solar cell at 50 MWp production line scale should be higher than 15.3% to assure a competitive operational cost for a new technology. Important to note that such estimations are very rough and do not take into account some issues related to implementation of low-cost non-vacuum processing technologies and advanced Si layers deposited technologies developed in frame of the ThinSi project.
Since efficiencies above 14% for Si wafer equivalent solar cells processed in frame of powder-to-substrate approach have been demonstrated, it can be concluded that the developed technological route is rather promising for advanced Si based PV.
The step-change has been realised by enabling high throughput processing using a set of advanced ceramics technologies, thereby avoiding costly crystallization and wafering steps. The substrates have been made from the low-cost Si powder. To fully exploit the advantages of the thin film silicon-based approach, the project has developed formation of high-quality poly-Si based layers on top of low-cost highly conductive Si the substrates. The new process consumes one order of magnitude less high-quality Si material than the conventional processing, since the main part of the solar cell structure consists of a low-cost Si powder based supporting substrate.
The global PV sector is characterized by rapid growth and requires further drastic reduction of PV system prices to compete with conventional energy sources. As it is stated in “A Strategical Research Agenda for Photovoltaic Solar Energy technology” , no clear technological “winners” or “losers” can be identified yet and many laboratory options still have to proof their potential for commercial production on a large scale. Therefore it is still important to develop of a portfolio of options and technologies rather than a limited set. It means that it is necessary to build a sufficiently broad European PV research sector and to allow new low cost and/or high efficiency options to show their long term potential.
According to this vision ThinSi was focused on impact of the Energy.2009.2.1.1 call “Accelerated market development of cost-effective and more efficient thin-film photovoltaics” as described below:
Reaching the target of cost effective manufacturing of thin Si based solar cells:
To realise a strategic Research Agenda developed in frame of the PV technology platform the cost effective manufacturing of thin Si based solar cells in frame of ThinSi project targeted the cost of PV modules well below 1€/Wp. This goal has been focused on the realization of the following options:
• Processing of low-cost Si based substrates using a “powder-to-substrate” concept.
This approach reduces the cost of the Si based solar cell substrate up to one order of magnitude compare to the cost of Si wafer, widely currently used in frame of wafer-based crystalline silicon technology. The following innovative technologies were developed for fabrication of low-cost substrates: (i) Spark plasma sintering, (ii) hot pressing, (iii) thermal spraying, (iv) exfoliation and bonding of thin Si layers on low-cost Si powder based substrates.
• Cost–effective deposition of high-quality active thin Si layers, which will serve as the base of thin-film solar cells.
This option has been realised by probing and optimization of several competing deposition methods, with the common general trend based on a low-temperature processing budget. Following this trend the following alternative versions for the cost effective deposition process has been tested: (i) the deposition temperature for epitaxial processes, below 1000 ºC, (ii) plasma assisted deposition of poly Si thin layers at middle range temperatures (600-700 ºC), (iii) solid phase crystallization of a-Si(a-Si:H) deposited at low (room temperature-200 ºC) on low-cost poly-Si substrates at 800 – 1000 ºC. The proposed realization is in line with the recommendations described in “A Strategical Research Agenda for Photovoltaic Solar Energy technology”, according to which the following topic have to be addressed: “processed and large area equipment for low-cost plasma deposition of micro/nanocrystalline silicon solar cells. The interplay plasma/devices/upscaling effects should be fully mastered”. It was demonstrated that the following, alternative to well established CVD based processes, innovative deposition technologies can be used: (i) high temperature magnetron sputtering; (ii) high temperature PECVD, (iii) thermal spray.
Thus, the overall cost reduction of hybrid solar cells at the step of the active thin Si layer formation is provided by: (i) low-Si material consumption (thin ~15-20 µm poly-Si layers), (ii) low-temperature budget for their deposition/crystallization.
• Optimization of methods for the low-temperature deposition of antireflection coatings based on oxides and nitrides as well as optimization of composition, quality and stability of such materials
This option has been realised by: (i) substitution of expensive ITO based ARCs by less expensive ZnO doped and FTO low-cost materials based layers, (ii) implementation of low-cost non-vacuum methods for the deposition of oxide based ARCs (ALD, ESAVD).
Implementation of all mentioned above options was based on “a better understanding of relationship between the deposition processes and parameters, the electrical and optical properties of the deposited materials, and the device properties and improvement of the quality and stability of transparent conductive oxides” as it is stated in the description of the content/scope of topic Energy.2009.2.1.1.
The main concept of ThinSi is to develop a solar cell process chain based on cost-effective manufacturing of thin Si based solar cells and modules. Thus, a technological platform, which includes innovative equipment and cost effective up-scalable processes has been created. The platform covers the complete fabrication process chain and potentially enables cost-effective and large scale production. Although not all technological routes tested in frame of the ThinSi project have been realised on the laboratory scale, the potential applicability of the basic innovative processes developed in frame of ThinSi project for fabrication of innovative solar cell structures have been demonstrated. As a result, ThinSi has created a general capability to establish a platform for the innovative cost effective processing of thin-film Si based solar cells and modules. Such a platform includes:
• A set of advanced equipment units, which provide establishment of a cost effective processing line for fabrication of thin-film Si based solar cells.
• Cost-effective low-cost solar cell fabrication process chain including quality monitoring
• Process integration documentation
• Device design and fabrication rules
Reaching the target of more efficient thin-film photovoltaics
Following the recommendations described in “A Strategical Research Agenda for Photovoltaic Solar Energy technology” , which stated that in order to strengthen the PV positions on the global market and to develop more efficient thin-film Si based PV, the following topics have to be addressed:
• Development of specific high-quality low cost transparent conductive oxides suitable for large area high performance (>12%) modules.
This option has been realised by development of non-vacuum process for large scale TCO deposition using ESAVD technology.
• Demonstration of higher efficiency thin-film Si devices (>15% on lab scale), improved understanding of interface and materials properties, of light trapping, and of fundamental limits of thin-film Si based materials and devices.
These option has been addressed by: (i) development of low-cost highly conductive oxides and high-quality passivation layers between TCO ARCs and Si substrates, similar to that what was realised and demonstrated by SANYO , (ii) development and utilization of embedded porous Si Bragg reflector.
Reaching the target of accelerated market development
ThinSi consortium has addressed issues concerning a possibility for potential commercialization on several levels:
• The manufacturers of low-cost Si powder feedstock as well as manufacturers of low-cost Si powder based substrates prepared by ceramic technology and hybrid Si wafer equivalent solar cell bases have the possibility of serving European, Asian and US customers.
• The manufacturers of the equipment for the cost-effective deposition of thin Si and TCO layers will have the possibility to serve customers worldwide with such equipment and also with the corresponded technologies, which can be realised using each concrete equipment unit.
• The manufacturer of the analytical equipment and procedures for the characterisation of electrical, composition and morphological properties of solar cell materials and solar cell structures have the possibility of serving customers worldwide.
• Large European companies tend to buy a complete set of technologies and processing equipment, rather than to develop processes and devices themselves. European Research Centres and SMEs involved in ThinSi are natural providers of such service after successful realization of the project goals.
Statements about commercialization potential by the SME and industry partners of ThinSi is shown below:
Commercialization potential for the SME hardware providers
In order for Europe to keep the actual growth rates of the PV sector (almost 50% p.a. over the past five years), the PV manufacturing industry needs new tools to realize the new concepts for the cost effective processing of solar cells. OIPT has designed equipment for scaling processes to production levels, aiming to build a closed circle of the Si based heterojunction solar cell processing chain, which can be realised in a cluster type system. OIPT is well positioned to bring the results of a successful project to market, by offering a qualified process and tool, including transition towards production customers. IMPT has developed and will bring to the European market a specialised equipment based on non vacuum ESAVD coating method for the fabrication of TCO layers for solar cells. NT-MDT has developed and will bring to Europen market special analytical equipment, which will allows to make monitoring of electrical and structural properties of individual solar cell materials and solar cell structures interfaces on nano-scale.
Commercialization potential for the SME technology providers:
PyroGenesis has developed and can bring to European market advanced non-vacuum technologies, which can be considered as a “core” of an advanced cost-effective PV: (i) thermal spray based technology for deposition of low-cost metallurgically grade Si powder to process cheap Si based supporting substrates, (ii) inert atmosphere plasma spraying based technology for deposition of electronically grade Si on the top of low-cost substrates, (iii) advanced thermal spay based technology for deposition of Al-based powder for the back side contact of solar cells
IMPT’s unique non vacuum ESAVD based coating technology is well positioned to offer a cost-effective TCO layers for solar cells and can contribute to bringing the successful project into market, by providing a qualified process and tool, as well as support for customers towards production.
OIPT in ThinSi has developed and therefore is able to bring to market innovative processes to provide a middle range temperature (600-700 º) in-situ crystallization of thin Si layers deposited on Si powder based substrates by magnetron sputtering or PECVD.
Commercialization potential for SME end-users:
The ThinSi end-user ISO, as well as European PV SMEs in general, have got a possibility to install a cost effective large scale processing of advanced thin-film Si based solar cells. Such installation does not require too many additional modifications of the existing wafer based or thin film technologies, since will combine advantages of both approaches and moreover will build a bridge between them. SMEs, which are dealing with the Si wafer-based crystalline silicon technologies will be able to reduce the PV cost essentially by switching the solar cell processing to the low-cost Si wafer equivalent + thin Si film based chain instead of the crystalline wafer rout widely used recently.
Commercialization potential for other participants:
All RTD partners are able to protect their technological developments by patents, since a number of advanced technological developments have been demonstrated in frame of the ThinSi activity. Several patent applications concerning thermal spray based processes are under preparation currently. RTD, as well as some SME partners are able to provide results of innovative developments to relevant SMEs and Large European companies, which are not involved in ThinSi.
Impact towards principal objectives of the ENERGY-theme
A key element of Theme 5 (Energy) is to provide research into development and demonstration of integrated technologies for electricity production from renewables. Research should increase overall conversion efficiency, cost efficiency, significantly drive down the cost of electricity production from renewable energy resources. To realise such goals research, which was performed in frame of ThinSi included developments and demonstrations of new processes for photovoltaic manufacturing, including the manufacturing of equipment for the PV industry. ThinSi contributed to this objective as it aimed to develop a complete innovative solar cell processing chain based on implementation of an innovative “powder-to-substrate” concept combined with the high throughput deposition of thin Si films on such substrates. The SME and industry partners in ThinSi has got an opportunity to transfer the new knowledge into their own products, as they see a market potential both as solar cells, technology and fabrication tool providers.
By realisation of the mentioned above developments, ThinSi has contributed to the principal objectives of the Energy-theme: to establish long term strategies for next generation photovoltaics (both high-efficiency and low-cost routs) and will improve the competitiveness of the European PV industry.
Steps which has been done to bring about expected impact
The most important objective in order to make high volume and cost effective fabrication of Si based solar cells in Europe is to solve the material and the processing costs issues. The consortium is confident that these issues have been addressed by the ThinSi project. In order to realise goals of the ThinSi project and provide their realisation, combination of well established processes with some innovative steps has been used: (i) processing of Si powder with a desirable purity, crystallinity and size, (ii) ceramics technology for the processing of Si powder based supporting substrates, (iii) high throughput deposition of thin Si layers for the cost effective formation of the solar cell base, (iv) well established conventional wafer based technology for the processing of solar cells and modules, for a comparison.
At the same time some innovative developments, to provide a breakthrough in the PV field in frame of the overall strategy of the ThinSi project have been realised: (i) new deposition and crystallization methods and the related equipment for high quality silicon active films at low-temperature budget on low-cost Si based substrates, (ii) effective technological solutions for the solar cell processing using a low-temperature heterojunction approach and advanced methods for optical confinement, (iii) up-scalable non -vacuum technique for formation of the TCO ARCs, (iv) A quality inspection tool and methodology for the analysis and control of solar cell structures and interfaces on the nano-scale.
As a result a new complete process chain, which potentially can substitute the conventional Si wafer based solar cell processing technology has been developed and disseminated. It is important that the results of the project, and the new capabilities of the partners, have been disseminated at conferences and exhibitions and were a part of the partners’ own strategies. Separate ThinSi dedicated work-shop "Advanced concepts for Si based PV", which was organised in collaboration with EU project nanoPV, R2M-Si and Hipersol has been an important tool in this respect.
Important to note that none of the individual partners or participating countries would be able to succeed in the development of the proposed technology new process flow based on a “powder-to-substrates” concept.
Dissemination activities and exploitation
It was supposed that the interest of the industrial partners and European industry in general will be protected by securing the IPR through patents. Since project time has been limited and all partners were overloaded by on-going developments, no patent applications were filed so far. Nevertheless, several patent applications are under preparation currently (thermal spray related). Patent applications will be submitted in the nearly future. Such situation prevented, partially, dissemination of some important results, since IPR have to be secured. After securing the IPR, the project partners will disseminate much more knowledge outside the Consortium, to allow other actors to contribute to technology development and deployment.
Confidential information provided by any of the partners was used freely within the scope and for the duration of the project. All information provided to the project team (or part thereof) was assumed to be 'Commercial in Confidence' and the provider loosed no rights through its disclosure to the project team.
To assure successful exploitation of the ThinSi scientific and technical results, dissemination and exploitation activities was running in parallel with the coordination of the project so that all activities were linked to successive knowledge transfer. Disseminating and exploiting ThinSi achievements were considered as an integral part of all workpackages and a responsibility for all partners. Integrity and consistency of all dissemination and exploitation efforts was supervised in WP5 - Dissemination and exploitation. The efforts allocated to these activities were take place along the following main lines:
1. Dissemination of results obtained: All partners were collaborating in the dissemination of the results among the research community, the photovoltaics community and their respective industries and markets in a broad sense, including other FP7 programme participants.
2. Exploitation: Industrial take-up of the technology developed in ThinSi played a central part in the project. To maximise the likelihood of success, relevant project partners prepared an “exploitation assistance package”. The package contains manufacturing procedures, infrastructure and IPR.
The planning of dissemination activities, which is a horizontal activity along the overall project lifecycle, was started immediately in the first project meeting. Dissemination policies were based on three major dissemination channels. Each dissemination activity was designed as a blend of dissemination activities from one or more channels, with respect to the target group(s). The three channels and their components (in bold) are:
a. Online dissemination: A project website provided a first access point for interested scientific and business parties into the ThinSi project. The objective of the website is to create a community of interested parties around the project, to accelerate their involvement and to create awareness of the results.
b. Non-electronic dissemination: Classical vehicles of knowledge transfer such as open project seminars intended for industrial end-users, articles in technical journals, peer-reviewed publications in scientific journals and presentations in conferences were focused on the dissemination of project results, mainly to experts and professionals.
c. Personal dissemination: As described in ThinSi DoW, each partner represented a network of contacts, technology and market overview that is much greater than the consortium itself. The scientists and engineers working for ThinSi were actively seeking to disseminate the results through contacts in formal networks (e.g. participation in other projects and involvement in organisations) and in their personal informal networks.
Plans for use of results
The ThinSi partners exploited the results in many directions. The research partners (SIN, ISE, IMEC and ENEA) were utilizing the academic potential in scientific publications and as basis for new research. They will also offer product and process development to industries. Two industrial partners, OIPT and IMPT, offered highly sophisticated production equipment (vacuum and non-vacuum techniques, respectively) to the global PV fabrication market. Another industrial partner, NTMD, offered sophisticated analysis equipment. ELKS offered specialised Si powder qualities for the new PV substrate technology. PYRG is going commercialise technology based on thermal spray of Si powder for processing of solar cells.
Exploitation for each partner
SINTEF: With the results from ThinSi, SINTEF is able to offer a comprehensive package of knowledge about low-cost Si substrates for solar cells to industrial companies on a world wide scale. Through industrial development projects, SINTEF offers assistance to the companies in setting up their own production facilities. SINTEF is also able to produce small series of substrates in our own laboratories. To develop SINTEF's capabilities further, SINTEF intends to continue the cooperation with the ThinSi partners in new research projects inside and outside the European framework programmes. In particular, main developments, which were realised in frame of ThinSi is transferred to new EU project application – Waste2PV (on a second stage evaluation at the EC) in which So powder based concept is a central line of the project activity. Pyrogenesis and IMEC are the partners in this new application coordinated by CEA (France). Moreover, Si powder-to substrate concept is a part of the development in a new EU project EuruSunMed, coordinated by CNRS (SINTEF is a partner). As a marketing tool for services and research capabilities SINTEF intends to publish the project’s results – after securing IPR – in high quality peer-reviewed journals, at relevant conferences, and in application-focused technical journals.
ISE Freiburg: High throughput silicon deposition and epitaxy is a key process for all c-Si thin-film solar cell approaches. Within ThinSi, ISE has optimized the ConCVD processes/tools, and adapted silicon epitaxy to a cell concept with best cost perspectives. ISE is planning to exploit these results by licensing the technology to industrial partners, who will be able to secure a quick ramp-up of a cell production. Small-series production of c-Si thin-film solar cells/wafer equivalents are possible also at ISE in its new SIMTEC lab. The equipment and processes will be used in accompanying and further R&D projects, also in continued projects with the ThinSi partners.
The work on advanced solar cell processes like SiC hetero emitters, epitaxial emitters or plasma texture not only has applications for the ThinSi approaches, but can be used for other silicon solar cell approaches as well. Depending on the status of each process, exploitation is planned in connecting public R&D projects and/or industrial projects. As usual for the Fraunhofer philosophy, close cooperation with industrial partners is planned to disseminate as much as possible of the know-how for industrial application.
In addition to the exploitation issues, ISE has disseminated the open/secured knowledge in publications and workshops, using the whole variety of publication channels.
IMEC: When the project is successfully finished, IMEC will be able to offer a comprehensive knowledge portfolio on thin film solar cell manufacturing on low-cost substrates to industrial companies on a world wide scale. The technique developed in this project nicely complements our existing expertise on thin film epitaxial Si cells on low-cost multi-crystalline substrates, and our knowledge on very thin film poly-crystalline cells on glass and ceramic substrates.
As a marketing tool for our services and research capabilities IMEC, likewise SINTEF intends to publish the project’s results – after securing IPR – in high quality peer-reviewed journals, at relevant conferences, and in application-focused technical journals.
To develop our capabilities further, we intend to continue the cooperation with the ThinSi partners in new research projects inside and outside the European framework programmes.
ENEA: Deeper understanding of growth specifics of layer for solar cell applications through optical characterization has been achieved. Methodology for optical test on single films and multilayer stacks has been elaborated with the stress on the complexity of mathematical modelling for such devices. The elaborated approaches are thought to be useful both for further exploitation in solar cell technology development and for similar applications (TCO-including multilayers). Hence further cooperation with the project partners is auspicious and will be fruitful thank to the basis for systematic and reliable characterization thought to be created within this project, in terms of both methodology and relevant optical benches construction. The scientific results of this research have been presented at high-level international conferences and in peer-reviewed journals.
University of Nottingham: The science and technology generated from the study of the new formulation and deposition of low-cost TCOs based on doped ZnO in ThinSi project allowed for UNOTT to secure new IPs in this field and transferred the knowledge and technology to SME such as IMPT for scaling up the ESAVD process for large area non vacuum deposition of low cost TCOs onto 8” Si wafers for solar cell applications.
In addition, UNOTT has been committed to the dissemination of non IP sensitive information to both academic and industry in UK and Europe via seminar, workshop, international conferences, annual report, and reputable peer reviewed journals. UNOTT also intends to continue to work with ThinSi partners in other new research projects inside and outside the European framework programmes.
Oxford Instruments Plasma Technology Ltd: OIPT can offer qualified tools and processes for at least two process steps in the fabrication of thin film PV devices. It has an established position as an international equipment supplier to both the R&D sector and to specific production niches, and is capable of growing with the opportunity this project presents. Its minimum expectation is to increase company turnover by 10% (from approximately 40M euro) by equipping a few PV manufacturers. There is clearly scope for this project to significantly impact the future of the business.
Elkem Solar AS: Elkem Solar produces Solar grade silicon through a multi-step metallurgical process route. In this route silicon fines side-streams of different quality are available. The primary goal of Elkem Solars participation in the project is to increase its knowledge of these side-streams and their potential as substrates towards thin film solar cells. Furthermore, Elkem Solar wishes expend its network in the area of thin film solar cells.
PyroGenesis S.A.:ThinSi project is of great importance for PyroGenesis from the exploitation of results point of view. It is an opportunity for Pyrogenesis to explore the application of thermal spray coatings in a new and rapidly developing sector, the sector of Solar Energy Materials. Thermal spray technologies are not applied for the moment for such applications on a commercial basis. Therefore, PyroGenesis is the first company worldwide, who has develop know-how on thermal spray coatings for PV applications. Exploitation plans for PyroGenesis include the commercialization of the successful coatings as results of the project. PyroGenesis will have a serious role in the co-production of the final products, together with the other partners of the project. The company wishes to participate in the commercial production, as far as thermal spray is concerned. PyroGenesis is also willing to examine the feasibility of the installation of a full production chain in Greece, either alone or in collaboration with the project’s partners or with new partners who could obtain rights on the final results.
NT-MDT Europe BV: NT-MDT will bring to the European market two unique tools for monitoring of electrical and structural properties of individual solar cell materials and interfaces. The first one is an analytical instrument based on Atomic Force Microscopy (AFM) and Spectral Optical methods and instruments for solar cell processing on nanoscale level. The second system is a combination of AFM and ellipsometry for the investigation of the physical characteristics thin films and interfaces of the collar elements.
IMPT Ltd.: After a successful implementation of ThinSi it is envisaged that the commercialisation through the sales of specialised equipment and/or licensing of the ESAVD process for the non-vacuum deposition of TCOs to solar cell industry will follow. This will open up significant opportunities and generate enormous benefits, and revenues to IMPT.
IMPT adopted the formulation and deposition of low-cost TCOs (e.g. FTO) developed by UNOTT and applied them onto its specialised scaled up ESAVD, and demonstrate the technical and commercial viability of the manufacture of such low-cost new TCOs by the non-vacuum ESAVD deposition onto large area substrates.
The ThinSi project provided a platform for IMPT to continue to collaborate with other partners to develop and apply the non vacuum ESAVD deposition method for the cost-effective fabrication of high quality thin and coatings for other engineering and device applications, within and outside the European framework programmes.
Isofoton S.A.: Isofoton considers the use of low cost silicon substrates as a requirement for the future of the photovoltaic energy. Within the ThinSi project, Isofoton accumulated an important knowledge about the results of the introduction of low purity substrates into laboratory scale processing lines. On the basis of such results and cost estimations, an evaluation of a new low-cost technology has been done at Isofoton.
List of Websites:
4.1.5 Project data
The ThinSi project has an open website available on http://www.sintef.no/Projectweb/ThinSi(si apre in una nuova finestra). On this website information about the project vision, goals and events are published. The coordinating organisation was SINTEF, and the coordinator was Alexander Ulyashin (alexander.g.ulyashin@sintef.no, +47 93002224). The logo of the project is shown below:
ThinSi Consortium
Beneficiary Number * Beneficiary name Beneficiary short name Country Contact person
1 (Coordinator) Stiftelsen SINTEF SIN Norway Alexander Ulyashin
alexander.g.ulyashin@sintef.no
2 Fraunhofer Institute for Solar Energy Systems ISE Germany Stefan Reber
stefan.reber@ise.fraunhofer.no
3 Interuniversitair Micro-Electronica centrum VZW IMEC Belgium Ivan Gordon
gordoni@imec.be
4 Ente per le Nuove tecnologie, l’Energia e l’Ambiente ENEA Italy Anna Sytchkova
annak@enea.it
5 University of Nottingham UNOTT UK Kwang-Leong Choy
Kwang-Leong.Choy@nottingham.ac.uk
6 Oxford Instruments Plasma Technology Ltd. OIPT UK Mike Cooke
Mike.COOKE@oxinst.com
7 Elkem Solar AS ELKS Norway Ronny Gløckner
Ronny.Gloeckner@elkem.no
8 PyroGenesis S.A. PYRG Greece Michalis Vardavoulias
mvardavoulias@pyrogenesis-sa.gr
9 NT-MDT Europe BV NTMD Netherlands Julia Alexeeva
ualexeeva@ntmdt.eu
10 Innovative Materials processing Technologies Ltd. IMPT UK Admin/finance:
Corinna Pinfold
Corinna.pinfold@imptl.com
Technical:
Keith Wells
keith@imptl.com
11 Isofoton S.A. ISOF Spain Miguel Vazquez
m.vazquez@isofoton.com
ThinSi is an ambitious project to develop a solar cell processing chain for high throughput, cost-effective manufacturing of thin film silicon based solar cells on low-cost silicon substrates. The Si powder based substrates are made on the basis of an innovative powder-to-substrate concept. In line with the current trends in PV, this project aims to reduce the cost of solar cells and modules compared to those made by the conventional wafer based approach. The solar cell structure, which was the main subject for all ThinSi solar cell fabrication related developments, is very similar to a conventional bulk crystalline Si solar cell. The "ThinSi" solar cell structures consists of a low-cost Si powder supporting substrate (processed within the "powder-to-substrate" approach) and an active high- quality layer on top, which serves as a solar cell base. Thus, thin Si film based, but at the same time "all silicon" structure substitutes the Si wafer. The final structure can be called also as a "Si wafer equivalent".
It has been demonstrated that advanced silicon based substrates can be processed from Si powder using following state-of-the-art innovative technologies: (i) Thermal spraying; (ii) Hot pressing; (iii) Spark plasma sintering (SPS); (iv) block casting growth using low-cost Si powder feedstock followed by sawing of the grown Si ingots. It was established that all Si substrates, processed in frame of the mentioned above innovative technologies, have good enough mechanical and electrical properties to be considered as a supporting part of the Si wafer equivalent structure.
In addition to well established chemical vapour deposition (CVD) process for deposition of thin (~20-30 µm) Si layers, a set of innovative lower cost processes have been developed to realise the new concept: (i) thermal spraying, (ii) magnetron sputtering and plasma enhanced chemical vapour deposition (PECVD). New tools for deposition of thin Si layers (active layers in "Si wafer equivalent" structure) have been developed. High temperature sputtering and PECVD tools have been developed for in-situ deposition of poly or crystalline Si layers. All tools are designed for large scale silicon deposition, with capability to process 156mm × 156mm Si wafers. Moreover, an advanced Electrostatic Spray Assisted Vapour Deposition (ESAVD) tool has been developed and offer a non-vacuum and cost-effective coating method for deposition of transparent conducting oxide films. AFM/Raman/Ellipsometer/SNOM combined measurement tool and relevant methodology for the analysis of solar cell structures on nano-scale have been developed.
Four approaches have been investigated to make thin Si layer based solar cells: (i) the ‘high-temperature approach’ based on implementation of diffused or epi- emitters, (ii) the ‘low-temperature approach’ (hetero-emitter), (iii) the ‘diffused emitter" combined with TCO, (iv) the exfoliation and bonding of epitaxial silicon foils on low-cost Si powder based substrates. The highest solar cell efficiency achieved on Si powder based substrates in frame of "high-temperature approach was 11.9%. By using the exfoliation and bonding of epitaxial silicon foils with an implemented Bragg reflector, free standing structures were obtained which could be processed to solar cells and bonded to low-cost Si powder based substrates with efficiencies over 14%. It is concluded that approach bases on exfoliation and bonding of epitaxial silicon foils is a promising advanced technology for prospective thin Si based solar cell processing on low-cost Si based substrates.
Project Context and Objectives:
4.1.2 Project context and objectives
Today, in many regions of the world, PV systems can produce electricity at less than twice the cost of electricity purchased from the grid. Indeed, parity has already been reached in Japan and is foreseen to be reached in 5-10 years in other regions all over the world.
The PV module market is dominated by crystalline silicon solar cells. It is important to note that about 40% of the silicon module cost is from the silicon wafers. Currently, the silicon feedstock demand for PV applications is enormous and with a growing PV industry the need for crystalline silicon will increase dramatically. Therefore, a major part of current research activity is concentrated on a search for alternative silicon based solar cell concepts with reduced consumption of high-purity silicon. In conventional Si wafer based solar cell technology, most of the Si material acts as a mechanical carrier for the solar cell structures. However, since most of the optical absorption in Si takes place in the upper 15-30 µm, it is sufficient to use only thin Si layers with thicknesses in this range. Indeed, if special optical confinement schemes are implemented, even thinner layers can be used. In general, thin-film photovoltaics are assumed to become a market dominating technology in the long term and any development in this field is extremely important for the EU industry.
ThinSi is an ambitious project to develop a solar cell processing chain for high throughput, cost-effective manufacturing of thin film silicon based solar cells on low-cost silicon substrates. The Si powder based substrates are made on the basis of an innovative powder-to-substrate concept. In line with the current trends in PV, this project aims to reduce the cost of solar cells and modules compared to those made by the conventional wafer based approach.
The solar cell structure, which was the main subject for all ThinSi solar cell fabrication related developments, is very similar to a conventional bulk crystalline Si solar cell. The "ThinSi" solar cell structures consists of a low-cost Si powder supporting substrate (processed within the "powder-to-substrate" approach) and an active high- quality layer on top, which serves as a solar cell base. Thus, thin Si film based, but at the same time "all silicon" structure substitutes the Si wafer. The final structure can be called also as a "Si wafer equivalent".
In addition to well established chemical vapour deposition (CVD) process for deposition of thin (~20-30 µm) Si layers, it was necessary to develop innovative lower cost processes in order to realise the new concept. The following state-of-the-art innovative technologies for processing of advanced silicon powder based substrates have been considered in frame of the ThinSi project using: (i) Thermal spraying; (ii) Hot pressing; (iii) Spark plasma sintering (SPS); (iv) Block casting growth using low-cost Si powder feedstock followed by sawing of the grown Si ingots. It was established that all Si substrates, processed in frame of the mentioned above innovative technologies, have good enough mechanical and electrical properties to be considered as a supporting part of the Si wafer equivalent structure.
New tools for the cost-effective deposition of thin Si layers (active layers in "Si wafer equivalent" structure) were under consideration in frame of the ThinSi project activity. In particular, high temperature sputtering and plasma enhanced chemical vapour deposition (PECVD) tools have been developed for in-situ deposition of poly or crystalline Si layers. All tools have been designed for large scale silicon deposition, with capability to process 156mm × 156mm Si wafers. Development of an advanced Electrostatic Spray Assisted Vapour Deposition (ESAVD) tool, which can offer a non-vacuum and cost-effective coating method for deposition of transparent conducting oxide films, has been considered as an important activity towards non-vacuum technology for the cost effective processing of Si based solar cells. AFM/Raman/Ellipsometer/SNOM combined measurement tool and relevant methodology for the analysis of solar cell structures on nano-scale were considered as an important input into ThinSi activity, since in many cases such analysis is required to investigate peculiarities of advanced solar cell structures.
Important to mention that all project developments, concerning innovative technologies and equipment prototypes were based on a requirement to scale up and transfer to production lines by the end of the project or in a realistic time in a future. New market opportunities for the SME and industrial partners were considered as an important issue, for both, production tool suppliers and for end-users of the technology.
One of the targets of the ThinSi activity was to reduce the thermal budget for deposition of high quality poly-Si thin layers, since the highly doped Si based substrates contain high concentrations of impurities. Normal deposition is usually done by epitaxial high temperature growth at 1000–1200ºC and the risk to contaminate the growing/ crystallized high quality Si active layer on such a substrate is rather high. Therefore, several approaches to address this issue have been considered: (i) Lower temperatures for the deposition of thin Si layers; (ii) Alternative to conventional CVD deposition methods for Si layers based on thermal spray, PECVD and magnetron sputtering; (iii) Gettering.
The ThinSi solar cell structure, being wafer based and based on silicon, is very similar to traditional cells. This will ensure a low acceptance threshold in the solar cell industry, which presently is 95 % based on crystalline Si4. In spite of these similarities, several essential bottlenecks and developments in low-cost process methods, and therefore in the whole process flow for the processing of poly-Si on low-quality Si substrates solar cell, must be addressed:
(i) Reduction of cost and simplification of processing of low-cost Si based substrates using a “powder-to-substrate” concept (bottleneck 1)
(ii) Methodology for cost-effective deposition of high quality active layers on low cost substrates, including Si base, emitter and TCO antireflection coating retaining a high cell efficiency (bottleneck 2)
(iii) Thorough understanding of the electronic properties of the deposited solar cell thin films and their interfaces as a function of deposition parameters (bottleneck 3)
(iv) Development of advanced methods for optical confinement (bottleneck 4)
(v) Implementation of “powder-to-substrate” concept based processes flow for the solar cell fabrication to existing industrial lines currently operating in frame of single Si wafer based approach (bottleneck 5).
The complexity of the solar cell device structure requires high-level expertise in modelling, design, materials technology and device processing. The project partners have been chosen so that all the required experience will be available to the project: silicon feedstock processing, ceramics technologies, deposition and crystallization of thin Si layers, deposition of TCO layers, and expertise on materials and interfaces.
ThinSi project has the following specific objectives:
Objective 1
• To develop innovative methods for fabrication of low-cost Si based substrates (elimination of bottleneck 1). The development contains:
o Fabrication of Si powder with the desired purity, crystallinity, conductivity (~0.01 Ωcm) and particle size (up to ~100 µm).
o Processing of low-cost Si substrates from metallurgical grade (mg) or metallurgical up-grade (umg) Si powder by ceramics technologies (tape casting and thermal spraying) with crystalline Si seeding grains with diameter ~100 µm.
o Post processing of Si powder based substrates prior to the formation of the thin Si solar cell bases.
o Manufacturing procedures for pilot scale processing of Si powder based substrates
Objective 2
• To develop a process flow and advanced equipment for cost-effective fabrication of thin Si solar cells on low-cost Si based substrates, and assembly into modules (elimination of bottleneck 2):
o Development and optimization of prototypes of high throughput advanced equipment for the cost effective processing of Si powder based supporting substrates and thin Si based solar cells on their basis
o Development and selection of cost effective processes, for the formation of the Si base, the emitter and the TCO based antireflection coatings (ARC), and methods for their stabilization
o Development of metallization schemes for large scale thin Si based solar cells
o Development of a process flow and cost effective processing of thin Si based solar cells with conversion efficiencies up to 15%.
o Assembly of the advanced solar cells into complete modules at a processing cost targeting ~1 Euro/Wp.
Objective 3
• To analyse the electronic properties of individual solar cell materials and their interfaces as well as the relationship between the deposition parameters and the device properties (elimination of bottleneck 3):
o Analysis of optical and electrical properties of individual solar cell structure layers
o Analysis of the relationship between the properties of individual layers and interfaces, and the final properties of solar cells
o Ab initio simulation of electronic structure of solar cell materials and interfaces in solar cell structures under investigation for some selected cases
o Development of the quality inspection specific equipment and methodology for the express analysis of solar cell structures on nano-scale
Objective 4
• To develop new methods for optical confinement (elimination of bottleneck 4):
o Development of an internal reflector in thin solar cells on low-cost substrates
o Development of advanced wet chemical and reactive ion etching (RIE) texturing methods for thin Si layers and low-cost substrates
o Application of plasmonic nano-particles enhanced light absorption in thin Si based solar cells
Objective 5
• To implement the “powder-to-substrate” concept based process flows to existing industrial lines (elimination of bottleneck 5).
o Development of innovative up-scaling technologies and equipment prototypes that will be transferred to pilot production lines by the end of the project targeting processing cost ~1 Euro/Wp.
o To create new market opportunities for the SME and industrial partners of the project
Project Results:
4.1.3 Main S&T results/foregrounds
WP1: Si powder synthesis and low-cost substrate processing
Since all ThinSi developments related to fabrication of innovative low-cost supporting substrates are based on implementation of Si powder, in WP1, Si powders with the appropriate purity and diameter have been synthesised and used for processing of the low-cost Si based supporting substrates. During WP1, a set of cost-effective processing technologies for fabrication of Si based supporting substrates has been developed. The developed processes can be considered as a background for an innovative “wafer equivalent” technology based on powder-to substrate concept.
The Si powder based substrates are made on the basis of an innovative powder-to-substrate concept. In line with the current trends in PV, this project aims to reduce the cost of solar cells and modules compared to those made by the conventional wafer based approach. The solar cell structure, which was the main subject for all ThinSi solar cell fabrication related developments, is very similar to a conventional bulk crystalline Si solar cell. The "ThinSi" solar cell structures consists of a low-cost Si powder supporting substrate (processed within the "powder-to-substrate" approach) and an active high- quality layer on top, which serves as a solar cell base. Thus, thin Si film based, but at the same time "all silicon" structure substitutes the Si wafer. The final structure can be called also as a "Si wafer equivalent".
In WP1 It has been demonstrated that advanced silicon based substrates can be processed from Si powder using following state-of-the-art innovative technologies: (i) Thermal spraying; (ii) Hot pressing; (iii) Spark plasma sintering (SPS); (iv) block casting growth using low-cost Si powder feedstock followed by sawing of the grown Si ingots. It was established that all Si substrates, processed in frame of the mentioned above innovative technologies, have good enough mechanical and electrical properties to be considered as a supporting part of the Si wafer equivalent structure.
Information from WP1 is used as a background for further studies in WP2- WP5.
Fabrication of Si powder with the desired purity, crystallinity and particle size
The main aim of this part of the project (WP1 related activity) was to identify and to develop processes, which can be used as a background for an innovative “wafer equivalent” technology based on powder-to substrate concept. Therefore essential part of the activity was concentrated on Fabrication of Si powder with the desired purity, crystallinity and particle size. Experience from the partners using the Si powder for thermal spraying, hot pressing and tape casting followed by a spark plasma sintering, showed that a special refinement and adjustment of the Si powder quality was needed for each Si wafer sintering method, both with regard to conductivity and particle size distribution. ELKEM Solar has therefore started work on optimising the powder production. ELKEM Solar has produced and delivered 9 different Si powder qualities (with particle diameters from ~10 to ~100 µm, un-doped and doped) for use in low cost substrates based on feedback from the consortium. Several routes were investigated to sintering tape casting silicon tapes at temperatures below the melting point of Si. It was established that a spark plasma sintering (SPS) as well as hot pressing of silicon powder with and without doping elements are promising methods to produce Si wafers with the desirable properties (dense structure and conductivity < 0.01 Ωcm). By use of this method, substrates which satisfied the requirements of the project have been produced. For the substrates both Al and B doped Si powder has been used. B and Al doping has been done by mechanically mixing the Si powder with B/Al at SINT, while B-doped Si powder also has been produced at ELKS.
ELKS has produced and delivered different Si powder qualities for use in low cost substrates based on feedback from the consortium. The powders are in general tailored based on the following learning points from consortium feedback:
- In terms of Fe and other metal elements in the silicon powder, ELKSM was looking for qualities between quality A (metallurgical grade Si) and ESS (Elkem Solar Silicon™, solar grade quality).
- Highly conductive B-doped Si powder, which is suitable for sintering of Si based substrates with resistivity of about 0.01 Ohm.cm and below.
- Thermal spraying techniques at Pyrogensis require PSD between 10 and 90 µm. To be sure that the powder does not hinder flow or block the feed to spraying equipment, ELKS has produced a 45 – 80 µm PSD range
- Hot pressing and SPS techniques require PSD < 120 µm. After having taken out 45 – 80 µm, 0 – 45 µm and 80 – 120 µm types of Si powders were fabricated at ELKS and delivered to SINTEF
Moreover, it has been established that for the spray pyrolysis based thin film production a PSD range of 20-30 µm was preferred. Some problems and difficulties occurred upon producing Si powder with such a tight PSD range in this size region. However, ELKS has managed to produce a 1 – 50 µm powder first and will has developed further separation methods for taking out the 20 – 30 µm particle size range from the initial Si powder mix. The production of such small particle sizes with this narrow size range is a true challenge. Finally, it was possible to produce 10 – 50 µm fines using special crushing and sieving methods.
Figure shows a sketch of the optimised production method for feed powder to making thin – film substrates using thermal spray or SPS methods:
Figure 3: Production method from bi-product to substrate powder
The method is optimized in terms of resistivity, particle size, production yield and chemical purity.
As an example Table 1 show powders with low impurity level produced at Elkem Solar. All samples are upgraded with a HF leaching process to remove impurities / contamination from crushing / grinding.
Table 1. Undoped upgraded Si powders.
Sample Treatment Fe
ppmw Al
ppmw Ca
ppmw Cr
ppmw Cu
ppmw Mn
ppmw Ti
ppmw Ni
ppmw P
ppmw
ENV450 Upgr. A+ 12.0 6.1 95 1.5 5.3 2.20 0.59 1.3 1.3
ENV451 Upgr. A+ 9.6 2.9 150 < 0.5 1.4 2.80 0.55 0.55 1.5
ENV452 Upgr. A+ 11.5 2.8 150 1.1 1.4 0.41 0.80 1.1 1.2
ENV448 Upgr. A+ 7.9 3.3 120 <0.5 1.8 0.38 0.50 NA 1.2
ENV456 Upgr. A 51 7.6 160 <0.5 5.1 0.92 0.86 0.69 1.5
ENV482 Upgr. A++ 8.6 2.6 75 5.5 13 1.4 0.27 1.6 1.6
From Table 1 it can be seen that the total impurity concentration somewhere between 150 and 170 ppmw (excluding oxygen).
Crystallinity of Si powder has been measured by micro-Raman. Results for Si powders A+75, A+40, A+10, D size1 and D size 2 can be seen in Figure 4.
Figure 4: micro-Raman spectra for different types of Si powders produced by Elkem Solar
It can be concluded that at least sub-surface region (~0.5 µm) of Si powder is not crystalline, although XRD measurements show that the main part of the powder is fully crystalline. This result shows that amorphous sub-surface powder regions can affect results of sintering and can cause problems to achieve highly doped Si substrates. Therefore such regions have to be re-crystallized upon the sintering.
Optimisation of processing conditions for fabrication of Si based supporting substrates by spark plasma sintering and hot pressing
After and in parallel with the successful developments related to fabrication of Si powders with the desirable properties, enormous efforts in ThinSi project were focused on processing of low cost Si powder based substrates with resistivity ~0.01 ohm•cm. It was established that the substrate quality obtained by tape casting followed by a solid phase sintering below the melting point of silicon was inadequate, too dusty and with too high resistance. It was concluded that stable substrates with satisfactory quality were very difficult to produce due to the difficulties with sintering of silicon. Therefore several alternative routes were investigated to improve the sintering of silicon tapes produced by tape casting. It was established that spark plasma sintering or hot pressing of silicon powder with and without doping elements are the most suitable methods to obtain dense and conductive Si powder based substrates. Moreover, conditions for the processing of Si powder based substrates by thermal spray method have been found.
Images of SPS based Si wafers are shown in Figure 5:
Figure 5: Spark plasma sintered silicon wafers (a) disc with d = 80 mm (b) disc with d = 200 mm.
It was established that to sinter good quality Si powder substrate a partial melting at the powder grains is required. This requirement causes some technical problems (melting of the central part of sintered wafer), which have been solved by optimisation of the SPS process. SEM image of the SPS Si powder based substrate is shown in Figure 6.
Figure 6: SEM images of two different locations on SPS Si powder sintered wafer surface
From Figure 6 it can be concluded that SPS Si powder based substrates have multi-Si like structure. Thus, Si powder based substrates processed by SPS method have structural properties, which are comparable with those for multi-Si material. Since the grain size in SPS material is much larger than the initial size of Si powder (< 100 µm), it can be concluded that partial melting of Si powder under SPS sintering conditions occurs.
It has been established also that in case of a low-cost Si powder (mg-Si) the SPS samples surface has depressions in which a very porous structure can be seen (Figure 7).
Figure 7: Depressions on the surface of SPS Si powder based sample
The depressions often contain distinct hexagonally shaped platelets. The particles are shown to contain chromium and are therefore probably CrSi2 which belong to the hexagonal crystal system.
Following this strategy (melting of Si powder), an ingot was made by melting and recrystallization of silicon powder (A+) mixed with boron providing the required resistivity of 0.01 ohm•cm. The experiment was done at SINT using a Crystalox DS-250 furnace. The material was melted in the furnace and slowly solidified with accurate control of the temperature profile and gas flow to guarantee ingot quality. The wafer cutting was done at Fraunhofer ISE using standard sawing techniques. 4-probe conductivity measurements were done giving an approximate resistivity of 0.006 ohm•cm. The wafers from this ingot were intensively used in WP4 for the solar cell processing.
In addition to mentioned above approaches, which includes melting of Si powder, hot pressing method, which was used at temperatures below melting point of Si has been tested as well. Si powder based substrates processed by hot pressing at high pressure and high temperature consist of a mix of small and very small crystallites (Figure 8).
Figure 8: Secondary electron image and orientation image micrograph (EBSD) for sample sintered by hot pressing.
From Figure 8 it can be seen that no twins or other Coincidence Site Lattice (CSL) boundaries observed, i.e. no recrystallization. No texture observed. Such hot pressed samples can be subjected to any wet chemical treatment prior any vacuum based process, since voids can't be seen on the surface of such sample.
Nevertheless, since sintering process can be done at different conditions, in some cases, if pressure and temperature are not optimized, hot pressed samples can contain some porous areas and application of wet chemical treatments for such samples is rather questionable.
It can be concluded that etching of hot pressed Si powder samples has some peculiarities, which depend on the sintering conditions. In general, wet chemical based treatments can be applied to such kind of Si powder based substrates.
Optimisation of processing conditions for fabrication of Si based supporting substrates made by thermal spray of Si powder
Several families of Si substrates made of low cost Si-based powders were produced. Some of them showed properties which fulfilled the selection criteria applied in the project: (i) low resistivity, below 0.01 ohm•cm, and (ii) acceptable consolidation and mechanical stability.
Production of free standing low resistivity (~0.01 ohm•cm) Si layers using thermal spray method has been performed using silicon powder size of 45-75 micron Quality D. The low resistivity was achieved by mixing in house, Si powder with Al and Boron dopants from 1% to 15%. The thermal spray technology used was Flame spray for higher production rates and cost effectiveness.
Free standing Si substrates with various dimensions from 50x65 mm2 to 156x156 mm2 and thicknesses of 300-700 µm have been processed (Figure 9).
The problems with the low mechanical stability and dusting from the substrates were overcome by applying a hotter thermal spray parameters and a post spray heat treatment protocol.
Figure 9: Si free standing low cost substrates with dimensions 50 x 65 and 156x156 mm2 processed by thermal spray
Important to note that Si substrates sintered by thermal spray after optimization of sintering conditions have crystalline structure, as measured by Raman (Figure 10).
Figure 10: Raman spectra of Si powder based wafer processed by thermal spray
Important to mention that the surface of Si powder based substrates sintered by thermal spray is rather rough (Figure 11a), which has some advantages for applications in solar cells, since such roughness provide dark colour and therefore good light trapping properties for such substrates.
Figure 11: (a) - Surface, and (b) - cross-section SEM images of the thermal spray sintered Si powder based wafer
At the same time, cross section SEM image shows that substrates include a lot of voids
(Figure 11b), which do not affect essentially mechanical properties of such substrates, since it seems that voids can block evolution of cracks in such structures.
Development of wet chemical methods for texturing and pre-treatment of low-cost Si based substrates prior to formation of high quality thin Si layers
Several types of SPS Si powder samples have been tested after CP4 (HF:HNO3) etch. Results can be seen in Figure 12.
Figure 12: Surface and cross section SEM images of SPS Si powder sample after CP4 (HF:HNO3) etch
From Figure 12 it can be seen that dense sintered Si regions and inclusions of not fully sintered agglomerated Si (potentially Si with contaminants, or dopants which is not fully reacted). Also cubical-like structures can be seen (right image) due to not fully sintered Si powder. As a result, formation of porous Si structures on such material could be difficult to realise.
Incorporation of porous Si layers into the Si supporting substrate
Porous Si etching in thick, 50 x50 mm2 multi-crystalline p+ substrates, made out of the A1 part of the ingot made of re-melted Si A+ powder (made by Sintef-ISE). The Substrates were first chemically polished to obtain a reasonably flat surface. A porous Si stack in the top part of the substrate was etched. Total reflectance measurements shows a clear reflection band around 800-900nm (theoretical reflection maximum at 950nm), which is a very good result for a first attempt of etching a porous Si Bragg reflector in an ‘unknown’ substrate. X-SEM image of the stack of porous Si layers is shown in Figure 13. This stack is creating an optical Bragg mirror etched structure.
Figure 13: X-SEM at top surface of SPS substrate after porous Si etching
From Figure 13 it can be seen that etching of a porous Si Bragg reflector is possible for the case of Si powder substrate subjected to a melting, which provides formation of a dense crystalline structure.
Results of the WP1 related activity can be summarised as follows:
• Optimisation of silicon powder production, including B-doping process to obtain highly doped Si powder based substrates
• Identification and demonstration of SPS and hot pressing methods for fabrication of dense and highly conductive silicon powder based substrates
• Identification of thermal spraying of silicon as a method to produce dense and highly conductive silicon substrates and thin layers
• Low cost wafers have been made from re-crystallisation of low purity silicon powder doped with Boron.
• Bragg reflectors based on porous Si structures have been etched into multi-Si wafers made from the ingot grown from low purity silicon powder doped with B.
WP2: Thin film processes
Si deposition by alternative methods
Epitaxial layers have been deposited at temperatures of 1050°C and 1000°C in the Atmospheric Pressure Chemical Vapour Deposition system RTCVD160 at the Fraunhofer ISE. Epitaxial layers with defect densities of < 105 cm-2 have been achieved at a temperature of 1050°C (Figure 14).
Figure 14: Sample surface after epitaxial deposition and treatment by “secco” etching. Defects like etch pits and stacking faults can be seen.
Both etch pit densities and stacking fault densities have been taken into account. Defect densities below 2.6•105 cm-2 were measured for epitaxial Si layers deposited at 1000°C.
OIPT has developed special tools, based on PECVD and magnetron sputtering processes, which allows deposition of Si based layers at temperatures up to 700 °C (PECVD) and 700 °C by magnetron sputtering. It has been demonstrated that deposition of Si layers on Si substrates at elevated temperatures (~700 °C) leads to a partial crystallization of such layers.
10-20 µm thick silicon films were deposited by pulsed DC magnetron sputtering with substrate table temperatures up to 690 °C on highly doped crystalline silicon and thermal oxide substrates. Deposition rates of silicon were measured by ellipsometry and lift-off and surface profile measurements. The a-Si samples were annealed at different table temperatures and for various times lengths. Raman spectroscopy and ellipsometry was utilized to confirm crystallinity of the silicon films. To demonstrate the possibility to use magnetron sputtering for the cost effective deposition of silicon layers on different substrates, two magnetron sputtering systems at Oxford Instruments have been used, the PlasmaPro System 100 and System 400.
The PlasmaPro System 100 is a single target system which can be configured with one pulsed DC, DC or RF magnetron up to 254 mm diameter. It is capable of loading circular wafers up to 200 mm diameter or 156 mm square. The target to substrate separation can be automatically varied from 11 cm to 19 cm while maintaining high vacuum, 3 × 10-7 Torr. RF bias on the substrate table allows plasma-assisted deposition and the capability to run pre-cleaning for greater adhesion and control of film structure and stoichiometry. The substrate temperature control is provided by an embedded heater, with a temperature range of 20 °C to 700 °C. Figure 15 shows a general overview of the PlasmaPro System 100.
Figure 15 A schematic of the PlasmaPro System 100.
The PlasmaPro System 400 can be configured with up to four magnetron targets to a maximum diameter of 200 mm. The magnetrons can be energised individually as either pulsed DC, DC or RF modes for batch or single-wafer processing. The single process chamber is partitioned into four sections, isolating the sources from each other without the high cost of clustering several single process chambers. A rotating shutter eliminates cross-contamination and enables target cleaning and sputtering stability before the wafers are exposed.
The system is capable of loading multiple substrates up to 200 mm round or 156 mm square. The wafer table may be operated in either static or rotating mode i.e. either with the wafer to be deposited placed under the energised target or continuously rotated through the deposition flux from the energised target. Uniformity masks can be fitted to optimise uniformity in rotating mode, or removed for high-rate deposition in static mode. Substrate RF bias offers plasma pre-cleaning and plasma-assisted deposition. The substrate table can be heated up to 300 °C. Figure 16 shows a general overview of the PlasmaPro System 400.
Figure 16 A schematic of the PlasmaPro System 400.
A major advantage of this system is that multiple materials can be deposited in one chamber in a single process. Multi-layer stacks are automated by process recipes in the software. The PlasmaPro System 400 provides the ability to sputter metals, semiconductors, as well as non-conducting materials with film thicknesses from sub nanometer up to many micrometers. It is well suited to the deposition of solar cell materials for which the backside reflector layer, silicon base and emitter layer, passivation layer and transparent conducting oxide layer can be deposited in this one chamber and in a single sequence.
Both sputtering systems are fitted with a wafer handler, either in the form of a single wafer load lock or batch cassette loading, since this minimises particle contamination and preserves the vacuum conditions of the process chamber.
Deposition process conditions: Two targets were used for the work in this paper, a lightly and a highly boron doped p-type silicon target, with a resistivity of 1 Ωcm and 0.01 Ωcm, respectively. Although both targets are conductive, pulsed DC mode was used. Pulsed DC differs from full DC by the addition of a reverse voltage bias oscillation with variable reverse time and pulse frequency. This reverse pulse is routinely used with dielectric films as it eliminates target arcing by attracting electrons in the plasma towards the target surface to maintain charge neutrality at the end of each deposition pulse. Highly doped silicon can be sputtered with full DC; however, pulsed DC is commonly used for silicon deposition to suppress arcs.
The process recipe includes four steps prior to deposition, a pump down step to 3 × 10-7 Torr to reduce contamination from atmospheric gas admitted during wafer loading, a dwell step for high temperature processes to enable the wafer to attain temperature equilibrium, an in-situ substrate clean using the table RF and a target clean (System 400 only). The DC power was set up to 3000 W during silicon deposition and was pulsed at 100 kHz for all runs with a reverse current time of 4 µs. For the majority of silicon samples, argon was the only process gas used and the process pressure was kept at 2 mTorr. This results in deposition of hydrogen-free silicon. However, the option to add hydrogen or other gases, for example phosphine and diborane for in-situ doping is feasible.
Thin silicon layers have been deposited with various table temperatures up to 690 °C on crystalline n-type silicon <100>, thermal oxide, aluminium and glass substrates for comparison. The table temperature was held to within 10 °C and the process parameters varied to optimize the silicon crystallinity. For a-Si deposition the deposition rate was optimized as a function of process parameters and target to substrate separation.
After deposition of silicon samples with various thicknesses, post-deposition annealing in vacuum at table temperatures up to 1000 °C for up to 3 hours was performed. The deposition rate was determined by lift-off and a surface step profile measurement (Ambios XP-300). Optical thickness and refractive index measurements were modelled using an ellipsometer (WVASE, J.A. Wollam). The fraction of crystalline structure was evaluated by means of micro-Raman spectroscopy measurements.
Results: ~1 µm thick silicon layers were deposited on crystalline n-type <100> silicon and thermal oxide with various table temperatures up to 690 °C. Figure 17 shows micro-Raman spectra for as-sputtered silicon at temperatures up to 690 °C. All the films exhibit crystallinity with small fractions of poly-crystalline silicon inclusions as shown by the shoulders in the range of 510-515 cm-1. It can be seen that silicon deposited on silicon substrates is more crystalline than that deposited onto thermal oxide.
Figure 17: Raman spectra for 1µm silicon deposited at 690 °C on thermal oxide (blue) and silicon (black) showing the crystalline phase.
Figure 18 shows Raman spectra of Si layers deposited on Si and glass substrates at 650 °C by PECVD.
Figure 18: Raman spectra of Si layers deposited by PECVD on Si and glass substrates at 650 °C.
From Figure 18 it can be seen that that almost complete crystallisation can be achieved upon in-situ PECVD deposition of Si layer on Si substrate at 650 °C.
Pyrogenesis has developed process for the deposition of thin (20-50 µm) layers by thermal spray using Solar grade Si powder. Cross section SEM image of such layer is shown in Figure 19.
Figure 19: Cross section SEM images of thermal sprayed Si /multi Si
From Figure 19 it can be seen that Si powder based layer deposietd on Si substrtes are rather dense and do not containe high concentraion of voids. Important to note also that such layers are fully crystalline ccording to Raman measuremnts.
Electrostatic Spray Assisted Vapour Deposition (ESAVD) of TCO layers
IMPT’s own the intellectual property of the eco-friendly and cost-effective non vacuum thin film deposition technique based on Electrostatic Spray Assisted Vapour Deposition (ESAVD). This method is an advanced version of the spray pyrolysis and a variant of the Chemical Vapour Deposition (CVD) process. ESAVD involves spraying atomised precursor droplets across an electric field where the charged droplets will undergo decomposition and chemical reactions in the vapour phase and deposit onto a heated substrate. The homogeneous and/or heterogeneous chemical reactions can be tailored to synthesise the desired films for various applications. ESAVD is a disruptive coating technology. A novel atmospheric deposition technique and tooling based on ESAVD has been developed for the deposition of transparent conducting oxides to be applied to thin film silicon solar cells. A low cost and non vacuum tool has been developed based on ESAVD for the production of transparent conducting oxides thin films. This new ESAVD tool provides a non vacuum and cost-effective coating method and process which might not be realised by the existing available tools.
A prototype scale-up ESAVD system (see Figure 20) has been researched and developed for this project as a deposition tool with the following features:
a) Non vacuum atmospheric system
b) Wafer load up to 200 mm round/square.
c) High temperature platen to accommodate up to 200 mm square Si wafer heating up to 600°C
d) Platen carrier equipment to integrate with R&D scale-up machinery
e) Motion stage equipment
f) Adjustable spray atomiser and wafer substrate working distance
g) Dedicated spray atomiser and assembly
h) Spray motion control software developed
i) High voltage device up to 20kV
j) Chemical liquid delivery system sufficient for dispensing precursors during large wafer deposition.
Figure 20: A prototype scale-up ESAVD system
The optical and electrical properties of the ITO film deposited by ESAVD are summarised as follows: sheet resistance – 47 Ω/•; resistivity – 4.3x10-4; carrier concentration 5.4x1020 e/cm3; mobility 27.3 cm2/Vs; optical transmission > 88%; refractive index – 1.8-2.0.
Figure 21: Hardware of the prototype scale-up ESAVD system for processing large area ITO onto Si wafers.
Results of WP2 related activity can be summarised as follows:
• OIPT developed advanced magnetron sputtering tools, which provide possibility to deposit crystalline or polycrystalline Si layers in an in-situ process at temperatures around 700°C.
• PyroGenesis has demonstrated production of thin Si coatings of 10-15 micron thickness using solar grade Si powder for the thermal spray deposition method.
• A prototype scale-up ESAVD system has been developed
• Highly conductive ITO layers as well as several alternative TCO films have been deposited by ESAVD and magnetron sputtering methods.
• OIPT has develop processes for deposition of thick Si layers (10µm), (poly)/crystalline emitter layers (100-300nm), and thin emitter layer (5-20nm) for integration into Si based solar cell structures.
WP3: Analysis and properties of thin layers and solar cell structures
Analysis of composition and electrical properties of individual Si based supporting
substrates and thin Si layers
Properties of Si supporting substrates and thin layers is systematically analysed by XPS, XRD, SEM, Raman, conductivity, SIMS.
Moreover, an integrated "AFM/Raman/Ellipsometer combined measurement tool" has been developed by NT-MDTin order to characterize Si based solar cell structures on nano-scale. The combined tool consists of two systems. The first is the integrated AFM/Raman system with the optional SNOM option, and another is the integrated AFM/Ellipsometer systems.
Integrated system is equipped with three switchable ‘laser’ inputs, which provide the possibilities of analysis on photo-voltaic specimen on a nanoscale with various wavelength lasers. This state-of-art system can be updated friendly with a new type of digital controller, which enables to perform ‘SKPM and SCM’ methods in-situ. These techniques provide the abilities of conductivity analysis on both ‘electric and dielectric’ surface solar cells and Photovoltaic in combined with Raman analysis.
AFM/Raman integrated system mainly consists of:
• Integral Base,
• 100x100x10um closed-loop scanner,
• Optical AFM measuring head with the inserted cantilever holders, which allows use the Hollow cantilever for SNOM operation,
• Raman Spectrum with three lasers options of 473nm, 532nm, and 633nm, and the other lasers of 405nm, 488nm, and 785nm can be easily installed as well,
• Laser Radiation input module,
• Detectors – CCD camera and PMT (updating standard CCD to EMCCD for faster Raman Mapping)
• Optical viewing system with 100x, 0.7 NA, objective
• SPM controllers (updating to the digital controller, which allows the SKPM/SCM measurement simultaneously for conducting and dielectric layers of solar cells)
This system provides the full AFM (Topography/morphology, conductivity, photo-current, potential) characterization on the solar cell (front surface and cross-section) together with the Raman characterization.
AFM/Ellipsometer integrated system
For integration of AFM with ellipsometry, a Woollam M-2000 spectroscopic ellipsometer was used by NT-MDT. The whole stage was designed in order to integrate with AFM system including to the ‘adapter stage and ARM support’, which enables ‘Full AFM measurements on the solar cell cross-section surface and Ellipsometer measurement from the aside under the AFM probe’ (Figure 22 a). The optical light indent/receipt tube with fine adjusted screws for optical alignment was modified and the final diameter of these two optical tubes was modified from 1mm down to 10um, shown in Figure 22b. Such modification allows us to analysis the multi-layers on nano-meter scale, especially for textural solar cell with eliminating the surface roughness effects.
Figure 22: a) top-view of the design for the integrated AFM/Ellipsometer system; and b) the indication of the modified optical indent tube
This final integrated system mainly consists of ‘AFM basement, Smena AFM head, AFM controllers and Ellipsometer equipped with separate controller and software’. Such integrated tool allows to complete AFM measurements and analysis on the cross-section surface of solar cells.
Analysis of properties of TCO and passivating layers
A special methodology for optical characterization of individual layers, interfaces and final solar cell structures, based on traditional spectrophotometry and spectral multi-angle ellipsometry, within which novel approaches both for spatially resolved measurement, and for modeling of solar cells multilayers is being developed. Passivation properties of thin layers will be analyzed by QssPC and μ-PCD methods for which the procedures have be already prepared.
Hydrogen related phenomena at ITO/a-Si:H/Si heterojunction solar cell interfaces have been studied as well as advanced ellipsometric characterization of single layers and solar cell structures was performed preliminarily for bulk and surface in homogeneity of ITO and AZO from ENEA.
Solar structures of the type ITO(~80nm)/p+-a:Si:H(~5nm)/polished n-Si (~250mm) have been analysed for their optical and geometrical parameters. The depth profile modelling of the optical constants has yielded important information on free-carriers distribution within the film thickness.
Ultrathin a-Si layers from OIPT: a vast set of PECVD-deposited a-Si layers analyzed provides a data base for the analysis of the correlation of the film properties with the deposition conditions.
Analysis of interface properties of solar cell structures
Analysis of interface properties of the solar cell structures has been performed using SEM and IV measurements. In some cases TO/Si interface has been studied by XPS and TEM. SSRM and SCM measurements have been used for the analytical investigations and control of solar cell structures on a nano-scale.
Interface properties of Solar cell structures processed using passivation by dielectic oxides
A set of solar cell structures processed on the bases of Si powder based substrates was investigated to clarify influence of interfaces on their properties.
Below a brief overview of interface properties of a typical solar cell stricture processed at ISE Freiburg is given.
Solar cell structure “Hc2c” : highly doped Cz substrate with ~ 2 µm p++ BSF and ~ 22 µm p-type base, no HCl-gettering, no texturing, diffused 80 Ohm/sq emitter without oxide passivation and directly evaporated Ti/Pd/Ag front metallization, double layer anitreflection (TiOx, MgFx), processed in the "ISE solar cell run", eta=10.6%, VOC=609mV, JSC=26.0mA/cm2 FF=67.4%
Interface of “Hc2c” solar cell structure is shown in Figure 23
Figure 23: TEM image of MgO/TiO/epi-Si/Si solar cell structure
From Figure 23 it can be concluded that at the interface between TiO passivating layer and Si a thin SiO2 layer is present, which is formed upon deposition of TiO oxide. It can be assumed that such layer can affect efficiency of solar cells, since passivation properties of TiO can be affected by such a layer. More detailed studies are required to clarify this interface related problem.
Interface properties of Solar cell structures containing TCO layers
Experimental: The Indium oxide (IO)-based films were deposited by the pyrosol method. The film-forming solution for ITO contained 0.2M InCl3 /3%SnCl4 /10M H2O in methanol, for IFO - 0.2M InCl3/0.15M NH4FHF/0.8M H2O in methanol. The ultrasonic nebulizer operating frequency was 2.64 MHz. The deposition temperature varied in the range of 390-490º C. Argon was used as a carrier gas.
Transmission electron microscopy (TEM) as well as energy filtered transmission electron microscopy (EFTEM), was performed using with a JEOL 2010F instrument, equipped with a GATAN energy filter on standard cross section specimens of the films.
N-type Cz Si wafers with a resistivity of 4.5 Ωcm and 390 µm in thickness were used for the SC processing. The wafers were textured on the front side. Fabrication of p+nn+ structure was performed by diffusion from deposited boron and phosphorous containing glasses at temperatures around 1000ºC. After the removal of glasses all samples were etched in a HF-HNO3 solution to make the emitter shallower. ITO thin (~80 nm) layers were deposited on the p+ and the IFO layers on n+ side. The ITO films had a sheet resistance of about 50 Ω/sq and the IFO films of about 30 Ω/sq. After TCO deposition, samples 6.5×6.5 cm2 in size were cut.
Figure 24 shows interface region between ITO and Si substrate
Figure 24: TEM images of ITO/Si interface for the case of ITO layer deposited at 400 ºC.
From Figure 24 it can be seen that a buffer layer between ITO and Si substrate is formed. The composition of this layer has to be investigated in details, which is not a subject of this work. Nevertheless it can be assumed that most probably this ~2 nm thick layer consists of silicon oxide, which is formed upon the ITO deposition at quite high temperatures in presence of water.
It can be concluded that at the interface between ITO as well as TiO layers and Si a thin SiO2 layer is present, which is formed upon deposition. Formation of such layers has to be taken into account for the analysis of the solar cells properties.
Analysis of solar cells and solar modules properties
Conversion efficiencies of solar cells were characterized and certificated by ISE. Spectral response, dark and illuminated IV measurements were already performed for solar cells processed within the consortium.
Results of WP3 related activity can be summarised as follows:
• Integrated prototype system AFM/RAMAN/Ellipsometry has been developed within WP3 by T-MDT. The tool provides the Raman, Topography, Photo-current, et al done on the same area simultaneously. Topography surface characterization in a larger Z scale and higher resolution as well all the detailed topography investigation, conductivity measurements, higher resolution (SSRM), all is possible with the tool. Currently demonstration measurements are in progress.
• Detailed optical and electrical analysis of the porous Si multilayer structure, made out of the re-melted Si A+ powder acting as reflector. A clear reflection band can be seen around 800-900 nm, by total reflection measurements. Very uniform multilayer structure also visible by cross section SEM.
• Detailed analysis (SEM) of a front surface texturization based on plasma self masking process. Detailed analysis of the surface and interface between active layer and reflector as well as top surface after texturization (SEM). Very good uniformity, low reflectivity at desired wavelengths and low Si removal down to 0.6um were measured repeatedly.
• Structures of the type ITO(~80nm)/p+-a:Si:H(~5nm)/polished n-Si (~250mm) have been analysed for their optical and geometrical parameters. The depth profile modelling of the optical constants has yielded important information on free-carrieris distribution within the film thickness.
• The SEM-EDX study of lowly and highly doped SPS substrates reveals that the sintered silicon contains structures with a different composition. The grains with weak EBSD quality rise difficulty in detecting the crystal structure. Those structures are crystalline but probably highly stressed. The orientation of the EBSD pattern remains the same within the same large structure. EDX reveals Si, B rich regions and Si, B, O rich regions. In addition, intra-grain defects/dislocations in the crystalline Si are revealed by the microscopy after secco etch, indication a poor crystal quality. In general it can be said that the lowly doped Si SPS substrate quality is low.
• SIMS analysis of the Si based supporting substrate, made of recrystallised A+ powder, have been performed for B, C and O levels presence in the substrate. It was found that the O level is around 2E18 at/cm3 or less, there seems to be a gradient of C present in the top 300nm near the surface; from then on the C concentration stabilises at around 9E17 at/cm3 (well above the C detection limit of 2E17 at/cm3). The high Boron concentration was confirmed (p+ substrate) at about 2E19 at/cm3.
• Reference Epi-wafer equivalent solar cell with efficiencies of 14.5% was reached on monocrystalline Si substrates and 13% efficiencies were reached for powder based substrates. FZ material processed in the same batch resulted in solar cell efficiencies of 17%.
WP4: Design, processing, characterization and testing of solar cells and modules
Different approaches have been investigated in the project to make efficient thin Si layer based solar cells. In the conventional high-temperature approach, processes for the deposition of thin silicon epitaxial layers were applied. The junction was formed either by conventional phosphorous diffusion or by epitaxial deposition of the emitter. The second approach was the low-temperature heterojunction approach. In this approach, an amorphous silicon or silicon carbide layer was used as emitter which allows processes at low temperatures (< 250°C) and provide an outstanding surface passivation. Higher open circuit voltages (VOC) than with conventional diffused emitters are possible. The so-called HIT structure (Heterojunction with Intrinsic Thin layer) was used to achieve the best performance. For the third approach, a phosphorous diffused emitter was combined with a transparent conducting oxide (TCO). Therefore, indium tin oxide (ITO) was deposited on top of the thin silicon layer based structure. The fourth approach was the exfoliation and bonding approach, which uses porous silicon layers to detach high-quality epitaxial silicon layers and bond them to low-cost silicon powder based Si substrates.
Process chains
High-temperature approach: In the high-temperature approach, the epitaxial deposition of a thin crystalline silicon layer is done at ISE in atmospheric pressure rapid thermal chemical vapor deposition reactors at temperatures between 1100°C and 1200°C. High-quality epitaxial layers can be processed. The emitter is formed by two different methods in the high-temperature approach, either by emitter epitaxy or by emitter diffusion. Several solar cell runs were performed using the high-temperature approach.
The solar cell processing is described by the following steps:
• HCl gas gettering on Si substrates
• Epitaxial deposition by APCVD (Atmospheric Pressure Chemical Vapour Deposition) of ~ 2.5 µm Back Surface Field (BSF) and ~ 30 µm base
• Diffused phosphorous emitter/epitaxy (sheet resistance: ~80 Ohm/sq)
• Al back contact (full area, ~ 3 µm thickness)
• Directly evaporated front grid (instead of a photolithographically defined grid)
• Light induced metal plating
• Double layer anti reflection coating (DL-ARC)
• Edge isolation
Prior to the epitaxy process step, the samples with porous Si shortly HF dipped (with HCl added), the others were cleaned by “IMEC clean”. By high-temperature epitaxy, a back surface field (BSF) with a thickness of ~2.5 µm and a epitaxial base with a thickness of ~ 30 µm was deposited, after a pre-annealing step in H2 to close and reorganize the pores.
The “First joint IMEC-ISE solar cell run” was done using low-cost crystalline Si substrates and reference substrates. The goal was to assess the gettering efficiency of porous Si layers and the gettering efficiency of HCl gas gettering and assess the upper level of the conversion efficiency of such structures. Solar cells on p+ 99.94% UMG-Si substrates from ISE were processed together with p+ multicrystalline and p+ monocrystalline Si reference substrates.
In the next run named “ISE solar cell run” first solar cells on Si powder based substrates were processed applying a simplified cell process. It means that basically the process chain used without the porous silicon formation step and the plasma texturing. Also, instead of photolithography, the front contacts were directly evaporated leading to higher shading due to higher metallization finger thickness. Wafers of the ThinSi ingot A1, which was crystallized from Elkem Solar Si A+ powder at SINTEF and cut into wafers at ISE, were used as Si powder based substrates.
In the next solar cell run (“Second joint IMEC-ISE solar cell run”), Si powder based substrates were used again. To prevent diffusion of detrimental metals from the Si powder based substrate into the epitaxially deposited active layer, the porous Si layers were included again in this run.
Table 2: Results of the first joint IMEC-ISE solar cell run
Sample No. Substrate material Cell size [mm2] VOC
[mV] JSC [mA/cm2] FF
[%] ηcell
[%]
172A UMG 616 28.0 77.5 13.4*
175A UMG 620 28.6 76.5 13.6*
S4a mc (reference) 48 x 48 611 29.0 74.8 13.3*
M3 mono (reference) 48 x 48 644 30.4 78.4 15.3*
* independently confirmed by ISE CalLab
Whisker growing was observed during the epitaxial deposition step (see Figure 25). It was due to the high metal content in the substrates and resulted in shunting of the p-n junction and therefore in poor cell results (conversion efficiencies between 0.4% and 4.3%).
Figure 25: Whiskers can be seen on the ingot A1 substrates after epitaxial deposition of Si (even with applied HCl gas gettering).
It can be concluded that the metal contents in the Si powder based substrates are too high, still after HCl gas gettering. Substrates with lower initial metal concentrations have to be used and additionally, a porous Si layer from IMEC has to be integrated which acts as a diffusion barrier and prevents diffusion of detrimental metals from the Si powder based substrate into the active layer. Porous Si layers were used in the Second joint IMEC-ISE solar cell run.
To prevent diffusion of detrimental metals from the Si powder based substrate into the epitaxially deposited active layer, the porous Si layers were included again in the “Second joint IMEC-ISE solar cell run”. An overview of the results is given in Table 3. The porous Si layer can be used to getter detrimental metals within the substrate. It was reported recently by IMEC that there is a gettering effect by porous silicon. The gettering effect of HCl gas gettering was also reported earlier, the latest results were published within the ThinSi project in the special issue “Advanced concepts for silicon PV”. HCl gas gettering as well as porous Si gettering were applied in this cell run.
Unfortunately, no statistically relevant results could be found in the groups of solar cells including HCl gas gettering and porous Si layer etching (groups B, E, and G). One reason for the bad statistics in the cell results is a high number of broken edges and samples and broken cells. This might be due to the inferior material quality. Nevertheless, it could be demonstrated that the texturing is beneficial to the short circuit current and therefore to the cell efficiency. The highest solar cell efficiency achieved on Si powder based substrates was 11.9%. All cells of group F showed efficiencies above 10%, which means that milestone M11 has been reached.
Table 3. Cell results of the Second joint IMEC-ISE solar cell run
Group Wafer ID HCl gettering Porous Si Reflector Texture JSC [mA/cm2] Cell Efficiency η
B 5-6-7-8 YES NO YES 22.8 – 27.5 0.0 – 9.4
E 17-18-19-20 NO YES NO 9.9 – 24.3 0.4 – 3.6
F 21-22-23-24 NO NO YES 29.3 – 29.6 11.0 – 11.9
G 25-26-27-28 NO YES YES 24.6 – 26.1 1.5 – 1.8
H 29-30-31-32 N NO NO 26.0 – 28.4 4.6 – 9.6
Low-temperature approach: In the low-temperature approach solar cell runs were performed using the heterojunction approach. The summarized results of the HIT and (for reference) diffused emitter cells are presented in Table 4. The average value of the short circuit current (JSC) is about 2 mA/cm2 higher for the cells with diffused emitter because of a more effective anti-reflection coating (SiNx instead of ITO used for the HIT cells). The current densities of both types of cells are typical for a base thickness of 25µm without texturing, coated with an ARC. The VOC values of the HIT cells show a clear gain of 25 mV compared to the VOCs of the 80 Ω/sq diffused emitter cells. An average FF of 80% is obtained for HIT cells. Because of a lithography step which is not optimized for non-textured samples, some adhesion problems occurred during the lift-off process of the metal after evaporation. Therefore, the fill factor (FF) values for the cells with standard diffused emitter are lower than would be expected. In previous experiments, no difference was observed in the FF values for samples with HIT structure and diffused emitter structure.
For thin Si layer based solar cells on reference monocrystalline Cz Si, it could be demonstrated that the a-Si:H heterojunction emitter with intrinsic thin layer approach is beneficial to the VOC resulting in higher cell efficiencies compared to cells with standard diffused emitter.
Table 4: Comparison of the HIT cell and diffused emitter cell results, processed at IMEC
No. JSC VOC FF Efficiency η
[mA/cm2] [mV] [%] [%]
HIT cells 26.8 655 80 14.1
Diffused emitter cells 28.9 630 67 12.3
Diffused emitter with TCO approach
There were also structures processed within the project which consisted of an epitaxial Si base layer on low-cost Si substrates with a conventional diffused emitter and a transparent conductive oxide (TCO) at the front, with the latter deposited at UNOTT & IMPT. Some of the samples were processed to solar cells which were characterized at IMEC. The following substrate materials were used for the sample structures: ingot A1 p+ (made from A+ powder), ISE UMG p+ (as low-cost reference), and Cz p+. The substrate size was 50 x 50 mm2 for each structure. It is demonstrated that the best solar cell efficiency could be as high as 9.7%, which is comparable with the conventional technology, currently available at ISE.
Exfoliation and bonding of epitaxial silicon foils approach
Another approach which was investigated in ThinSi is the technique to exfoliate a thin silicon layer and bond it to a low-cost substrate. A high-quality monocrystalline silicon substrate can be used to process a so-called ‘separation layer’ by electrochemical anodization. Porosification is performed in such a way that a low-porosity layer and a high-porosity layer are created. After annealing in H2, reorganization of the pores occurs. The back surface field (BSF), the emitter, and the base can be deposited in-situ. A complete structure is presented in Figure 26.
Figure 26: Substrate with separation layer, low-porosity layer, and epitaxial cell structure for exfoliation.
After the separation layer (weak layer) formation and the epitaxial growth processes, the front surface processing is done while the epitaxial structure is still attached to the substrate. Afterwards, the structure is detached and the rear-side processing is done on the free standing foil. All steps of the exfoliation and bonding approach are shown in Figure 27.
Figure 27: Exfoliation and bonding of epitaxial silicon foils on low-cost substrates
Additionally, in this approach light confinement can be applied in terms of a bragg reflector, which means a stack of porous silicon layers of alternating low and high porosity. This feature leads to light trapping and thus an increased short circuit current JSC.
All process steps concerning porous silicon layer formation, exfoliation and bonding processes were performed at IMEC. The first solar cell batch of epitaxial foils in ThinSi was processed without Bragg reflector. The thickness of the active base was ~ 28 µm. The results of the free standing epitaxial foils before and after bonding are presented in Table 5. The best cells show efficiencies close to 12%. For the bonding on the various substrates, Ag based glue was used. There were some problems observed: Shunts could be observed on the edges of the cells, which are due to the liquid Ag based glue. That is why the cells were diced out after bonding. Also, handling problems of the free standing foils made the processing difficult.
Table 5: Results of free standing and bonded foils without Bragg reflector epitaxial foils
Device JSC VOC FF Eta
[mA/cm2] [mV] [%] [%]
Free standing cells I_free 19.0 621 73 8.6
II_free 25.4 628 75 12.0
III_free 24.2 626 75 11.4
IV_free 23.8 627 77 11.5
Bonded cells
I_bond 18.6 610 72 8.1
II_bond 25.6 627 75 11.9
III_bond Broken
IV_bond 24.2 626 75 11.4
For the second batch ‘advanced epifoils’ were created including a Bragg reflector and a separation layer, both prepared by electrochemical anodization. The thickness of the active base was 35 µm. The cells were diced out for measuring to minimize losses caused by shunts.
In Table 6 the results of advanced epifoils are presented before and after bonding. After bonding an increase of the JSC was observed, which is due to light trapping. But at the same time a slight decrease of the VOC occurred. The best efficiency of a free standing epifoil was 14.5%, whereas the best efficiency after bonding was 14.2%. The best efficiency of an epifoil bonded on the sintered substrate fabricated during the ThinSi project is 14.0%. This is close to the value which is requested in milestone M8 (~15%).
Table 6: Results of advanced epitaxial foils with included Bragg reflector before and after bonding
JSC VOC FF Efficiency η
[mA/cm2] [mV] [%] [%]
Epifoil 31.5 615 75 14.5
Epifoil on low-cost multi-c substrate 31.6 614 73 14.2
Epifoil 30.0 618 76 14.1
Epifoil on sintered substrate 30.3 617 75 14.0
Finally, mini-module was fabricated from thin silicon based solar cells, processed in frame of the project. Low-cost 99.97% UMG silicon substrates with size 100x100 mm2 were used. The efficiency of the active area of the mini-module is ~7.9%, which is well below the requirements for Si wafer/wafer equivalent based PV. It can be concluded that the large scale epitaxial growth processes still have to be improved and optimized, prior the implementation of wafer equivalent technology to industrial lines. It has to be noted, however, that all laboratory scale developments demonstrate that the advanced technology under the study is promising and needs further efforts for the up-scaling.
WP5: Transfer to production lines
In this report we present a cost model developed to analysis the industrial feasibility of low-cost substrate technology. The standard technology cost model is shown because of it is used as reference. General hypotheses used to build the cost model are presented. The cost model results of standard BSF-Al technology are presented and the industrial feasibility of this technology is analyzed. The concepts “return of investment” and the “annual relative profits” are introduced and will be used to analyze the industrial feasibility of the technology.
The process flow for standard technology and low-cost substrate approach are shown. The cost model results for low-cost substrate technology are presented and the industrial feasibility analysis is described. Efficiency thresholds at mass production level to balance the operational cost, the return of investment and the annual profits of standard technology using low-cost substrate technology are found. The cost model results for low cost substrate technology are shown in the following figure. We present three different substrate costs (30%, 50% and 70% of standard Cz-Si, that is 1.22 €/piece) and the standard BSF-Al technology, as reference; we show the sensibility of operational cost at 50 MWp production line scale versus average solar cell efficiency (Figure 28).
Figure 28. Operational cost for PV module based on low cost substrate technology at 50 MWp production line scenario. Three different substrate costs have been analyzed (30%, 50% and 70% of Cz-Si wafer cost) and the sensibility versus solar cell efficiency is shown. The operation cost of standard BSF-Al technology is shown as reference (green).
When substrate cost is 50% of Cz-Si wafer, the solar cell efficiency should be higher than 16.5% to assure a competitive operational cost; and when substrate cost is 30% of Cz-Si wafer, the efficiency solar cell at 50 MWp production line scale should be higher than 15.3% to assure a competitive operational cost for technology.
An estimation of exploitation cost for 500 MWp – 1 GWp factory scale have been done (Figure 29). For the BSF-Al standard technology, the operational cost for PV module is 0.58 €/Wp, assuming 17.5% as average solar cell efficiency. To balance this operational cost and this solar cell efficiency level using LCS technology is needed a 62% substrate cost of standard Cz-Si wafer for 1 GWp scenario.
Figure 29. Operational cost for PV module based on low cost substrate technology at 500 MWp – 1 GWp factory scenario. Three different substrate costs have been analyzed (30%, 50% and 70% of Cz-Si wafer cost) and the sensibility versus solar cell efficiency is shown. The operation cost of standard BSF-Al technology is shown as reference (green).
When substrate cost is 50% of Cz-Si, the solar cell efficiency should be higher than 16.8% to assure a competitive operational cost; and when substrate cost is 30% of Cz-Si wafer, the efficiency solar cell at 1 GWp production line scale should be higher than 15.5% to assure a competitive operational cost for LCS technology.
To estimate and to verify compatibility of the developed innovative processes on an industrial level, some technological developments have been performed at ISO in collaboration with IPMT and OIPT. In particular, it was demonstrated that ITO as well as a-Si:H layers can be deposited by magnetron sputtering on 156x156 mm2 Si wafers provided by Isofoton with the sufficient for industrial processes homogeneity.
Some attempts have been done to process large scale solar cells with ITO antireflection coatings at Isofoton. It has been concluded that screen printing metallization currently available at Isofoton, has to be modified in order to be compatible with the TCO based antireflection coatings, which are present in the most versions of the advanced solar cell structures developed in frame of the ThinSi project. In particular, the firing temperature has to be reduced up to ~400 ⁰C from currently available ~800 ⁰C. Development of a special metallization process on industrial scale requires additional efforts, time and resources, which were limited and not sufficient in frame of ThinSi.
Since large scale epitaxial based solar cell structures processed at ISE did not show sufficiently good efficiencies, it was concluded that optimisation of the basic processes for epitaxial deposition of thin Si layers still has to be done prior the testing of such structures on a pilot scale/industrial level.
Potential Impact:
4.1.4 Potential impact
Strategic impact
The impact of ThinSi conforms to the very ambitious targets that have been adopted for renewable energy in Europe, for example:
• In 2001, EU aimed to have 21% of its electricity coming from renewable energy sources by 2010 .
• In March 2007, the Heads of States and Governments of the 27 EU Member States adopted a binding target of 20% renewable energy from final energy consumption by 2020 .
• In September 2008, the European Photovoltaic Industry Association declared that photovoltaic energy could provide 12% of European electricity by 20202.
Although reliable PV systems are already commercially available and widely deployed, the cost of PV generated electricity is still too high to compete with electricity from non-renewable sources. Therefore, further development of PV technology with the aim to drastically reduce the turn-key system prices and cost/m2 cell area is crucial. Indeed, it was emphasised already in 2005 by the Photovoltaic Technology Research Advisory Council (PV TRAC) that such a development is possible.
For conventional silicon PV technology experience has shown that the price is reduced by 20% each time there is a doubling of installed capacity. As a consequence of this rule, a huge and heavily subsidised capacity would have to be installed if the required cost reduction should be realised by the economy of scale alone. Therefore a step-change in the PV performance is required.
ThinSi project demonstrated a complete process chain for cost effective processing of solar cells based on a “powder-to-substrate” approach. Using a cost model, developed by Isofoton it is estimated that when substrate cost is 30% of Cz-Si wafer, the efficiency solar cell at 50 MWp production line scale should be higher than 15.3% to assure a competitive operational cost for a new technology. Important to note that such estimations are very rough and do not take into account some issues related to implementation of low-cost non-vacuum processing technologies and advanced Si layers deposited technologies developed in frame of the ThinSi project.
Since efficiencies above 14% for Si wafer equivalent solar cells processed in frame of powder-to-substrate approach have been demonstrated, it can be concluded that the developed technological route is rather promising for advanced Si based PV.
The step-change has been realised by enabling high throughput processing using a set of advanced ceramics technologies, thereby avoiding costly crystallization and wafering steps. The substrates have been made from the low-cost Si powder. To fully exploit the advantages of the thin film silicon-based approach, the project has developed formation of high-quality poly-Si based layers on top of low-cost highly conductive Si the substrates. The new process consumes one order of magnitude less high-quality Si material than the conventional processing, since the main part of the solar cell structure consists of a low-cost Si powder based supporting substrate.
The global PV sector is characterized by rapid growth and requires further drastic reduction of PV system prices to compete with conventional energy sources. As it is stated in “A Strategical Research Agenda for Photovoltaic Solar Energy technology” , no clear technological “winners” or “losers” can be identified yet and many laboratory options still have to proof their potential for commercial production on a large scale. Therefore it is still important to develop of a portfolio of options and technologies rather than a limited set. It means that it is necessary to build a sufficiently broad European PV research sector and to allow new low cost and/or high efficiency options to show their long term potential.
According to this vision ThinSi was focused on impact of the Energy.2009.2.1.1 call “Accelerated market development of cost-effective and more efficient thin-film photovoltaics” as described below:
Reaching the target of cost effective manufacturing of thin Si based solar cells:
To realise a strategic Research Agenda developed in frame of the PV technology platform the cost effective manufacturing of thin Si based solar cells in frame of ThinSi project targeted the cost of PV modules well below 1€/Wp. This goal has been focused on the realization of the following options:
• Processing of low-cost Si based substrates using a “powder-to-substrate” concept.
This approach reduces the cost of the Si based solar cell substrate up to one order of magnitude compare to the cost of Si wafer, widely currently used in frame of wafer-based crystalline silicon technology. The following innovative technologies were developed for fabrication of low-cost substrates: (i) Spark plasma sintering, (ii) hot pressing, (iii) thermal spraying, (iv) exfoliation and bonding of thin Si layers on low-cost Si powder based substrates.
• Cost–effective deposition of high-quality active thin Si layers, which will serve as the base of thin-film solar cells.
This option has been realised by probing and optimization of several competing deposition methods, with the common general trend based on a low-temperature processing budget. Following this trend the following alternative versions for the cost effective deposition process has been tested: (i) the deposition temperature for epitaxial processes, below 1000 ºC, (ii) plasma assisted deposition of poly Si thin layers at middle range temperatures (600-700 ºC), (iii) solid phase crystallization of a-Si(a-Si:H) deposited at low (room temperature-200 ºC) on low-cost poly-Si substrates at 800 – 1000 ºC. The proposed realization is in line with the recommendations described in “A Strategical Research Agenda for Photovoltaic Solar Energy technology”, according to which the following topic have to be addressed: “processed and large area equipment for low-cost plasma deposition of micro/nanocrystalline silicon solar cells. The interplay plasma/devices/upscaling effects should be fully mastered”. It was demonstrated that the following, alternative to well established CVD based processes, innovative deposition technologies can be used: (i) high temperature magnetron sputtering; (ii) high temperature PECVD, (iii) thermal spray.
Thus, the overall cost reduction of hybrid solar cells at the step of the active thin Si layer formation is provided by: (i) low-Si material consumption (thin ~15-20 µm poly-Si layers), (ii) low-temperature budget for their deposition/crystallization.
• Optimization of methods for the low-temperature deposition of antireflection coatings based on oxides and nitrides as well as optimization of composition, quality and stability of such materials
This option has been realised by: (i) substitution of expensive ITO based ARCs by less expensive ZnO doped and FTO low-cost materials based layers, (ii) implementation of low-cost non-vacuum methods for the deposition of oxide based ARCs (ALD, ESAVD).
Implementation of all mentioned above options was based on “a better understanding of relationship between the deposition processes and parameters, the electrical and optical properties of the deposited materials, and the device properties and improvement of the quality and stability of transparent conductive oxides” as it is stated in the description of the content/scope of topic Energy.2009.2.1.1.
The main concept of ThinSi is to develop a solar cell process chain based on cost-effective manufacturing of thin Si based solar cells and modules. Thus, a technological platform, which includes innovative equipment and cost effective up-scalable processes has been created. The platform covers the complete fabrication process chain and potentially enables cost-effective and large scale production. Although not all technological routes tested in frame of the ThinSi project have been realised on the laboratory scale, the potential applicability of the basic innovative processes developed in frame of ThinSi project for fabrication of innovative solar cell structures have been demonstrated. As a result, ThinSi has created a general capability to establish a platform for the innovative cost effective processing of thin-film Si based solar cells and modules. Such a platform includes:
• A set of advanced equipment units, which provide establishment of a cost effective processing line for fabrication of thin-film Si based solar cells.
• Cost-effective low-cost solar cell fabrication process chain including quality monitoring
• Process integration documentation
• Device design and fabrication rules
Reaching the target of more efficient thin-film photovoltaics
Following the recommendations described in “A Strategical Research Agenda for Photovoltaic Solar Energy technology” , which stated that in order to strengthen the PV positions on the global market and to develop more efficient thin-film Si based PV, the following topics have to be addressed:
• Development of specific high-quality low cost transparent conductive oxides suitable for large area high performance (>12%) modules.
This option has been realised by development of non-vacuum process for large scale TCO deposition using ESAVD technology.
• Demonstration of higher efficiency thin-film Si devices (>15% on lab scale), improved understanding of interface and materials properties, of light trapping, and of fundamental limits of thin-film Si based materials and devices.
These option has been addressed by: (i) development of low-cost highly conductive oxides and high-quality passivation layers between TCO ARCs and Si substrates, similar to that what was realised and demonstrated by SANYO , (ii) development and utilization of embedded porous Si Bragg reflector.
Reaching the target of accelerated market development
ThinSi consortium has addressed issues concerning a possibility for potential commercialization on several levels:
• The manufacturers of low-cost Si powder feedstock as well as manufacturers of low-cost Si powder based substrates prepared by ceramic technology and hybrid Si wafer equivalent solar cell bases have the possibility of serving European, Asian and US customers.
• The manufacturers of the equipment for the cost-effective deposition of thin Si and TCO layers will have the possibility to serve customers worldwide with such equipment and also with the corresponded technologies, which can be realised using each concrete equipment unit.
• The manufacturer of the analytical equipment and procedures for the characterisation of electrical, composition and morphological properties of solar cell materials and solar cell structures have the possibility of serving customers worldwide.
• Large European companies tend to buy a complete set of technologies and processing equipment, rather than to develop processes and devices themselves. European Research Centres and SMEs involved in ThinSi are natural providers of such service after successful realization of the project goals.
Statements about commercialization potential by the SME and industry partners of ThinSi is shown below:
Commercialization potential for the SME hardware providers
In order for Europe to keep the actual growth rates of the PV sector (almost 50% p.a. over the past five years), the PV manufacturing industry needs new tools to realize the new concepts for the cost effective processing of solar cells. OIPT has designed equipment for scaling processes to production levels, aiming to build a closed circle of the Si based heterojunction solar cell processing chain, which can be realised in a cluster type system. OIPT is well positioned to bring the results of a successful project to market, by offering a qualified process and tool, including transition towards production customers. IMPT has developed and will bring to the European market a specialised equipment based on non vacuum ESAVD coating method for the fabrication of TCO layers for solar cells. NT-MDT has developed and will bring to Europen market special analytical equipment, which will allows to make monitoring of electrical and structural properties of individual solar cell materials and solar cell structures interfaces on nano-scale.
Commercialization potential for the SME technology providers:
PyroGenesis has developed and can bring to European market advanced non-vacuum technologies, which can be considered as a “core” of an advanced cost-effective PV: (i) thermal spray based technology for deposition of low-cost metallurgically grade Si powder to process cheap Si based supporting substrates, (ii) inert atmosphere plasma spraying based technology for deposition of electronically grade Si on the top of low-cost substrates, (iii) advanced thermal spay based technology for deposition of Al-based powder for the back side contact of solar cells
IMPT’s unique non vacuum ESAVD based coating technology is well positioned to offer a cost-effective TCO layers for solar cells and can contribute to bringing the successful project into market, by providing a qualified process and tool, as well as support for customers towards production.
OIPT in ThinSi has developed and therefore is able to bring to market innovative processes to provide a middle range temperature (600-700 º) in-situ crystallization of thin Si layers deposited on Si powder based substrates by magnetron sputtering or PECVD.
Commercialization potential for SME end-users:
The ThinSi end-user ISO, as well as European PV SMEs in general, have got a possibility to install a cost effective large scale processing of advanced thin-film Si based solar cells. Such installation does not require too many additional modifications of the existing wafer based or thin film technologies, since will combine advantages of both approaches and moreover will build a bridge between them. SMEs, which are dealing with the Si wafer-based crystalline silicon technologies will be able to reduce the PV cost essentially by switching the solar cell processing to the low-cost Si wafer equivalent + thin Si film based chain instead of the crystalline wafer rout widely used recently.
Commercialization potential for other participants:
All RTD partners are able to protect their technological developments by patents, since a number of advanced technological developments have been demonstrated in frame of the ThinSi activity. Several patent applications concerning thermal spray based processes are under preparation currently. RTD, as well as some SME partners are able to provide results of innovative developments to relevant SMEs and Large European companies, which are not involved in ThinSi.
Impact towards principal objectives of the ENERGY-theme
A key element of Theme 5 (Energy) is to provide research into development and demonstration of integrated technologies for electricity production from renewables. Research should increase overall conversion efficiency, cost efficiency, significantly drive down the cost of electricity production from renewable energy resources. To realise such goals research, which was performed in frame of ThinSi included developments and demonstrations of new processes for photovoltaic manufacturing, including the manufacturing of equipment for the PV industry. ThinSi contributed to this objective as it aimed to develop a complete innovative solar cell processing chain based on implementation of an innovative “powder-to-substrate” concept combined with the high throughput deposition of thin Si films on such substrates. The SME and industry partners in ThinSi has got an opportunity to transfer the new knowledge into their own products, as they see a market potential both as solar cells, technology and fabrication tool providers.
By realisation of the mentioned above developments, ThinSi has contributed to the principal objectives of the Energy-theme: to establish long term strategies for next generation photovoltaics (both high-efficiency and low-cost routs) and will improve the competitiveness of the European PV industry.
Steps which has been done to bring about expected impact
The most important objective in order to make high volume and cost effective fabrication of Si based solar cells in Europe is to solve the material and the processing costs issues. The consortium is confident that these issues have been addressed by the ThinSi project. In order to realise goals of the ThinSi project and provide their realisation, combination of well established processes with some innovative steps has been used: (i) processing of Si powder with a desirable purity, crystallinity and size, (ii) ceramics technology for the processing of Si powder based supporting substrates, (iii) high throughput deposition of thin Si layers for the cost effective formation of the solar cell base, (iv) well established conventional wafer based technology for the processing of solar cells and modules, for a comparison.
At the same time some innovative developments, to provide a breakthrough in the PV field in frame of the overall strategy of the ThinSi project have been realised: (i) new deposition and crystallization methods and the related equipment for high quality silicon active films at low-temperature budget on low-cost Si based substrates, (ii) effective technological solutions for the solar cell processing using a low-temperature heterojunction approach and advanced methods for optical confinement, (iii) up-scalable non -vacuum technique for formation of the TCO ARCs, (iv) A quality inspection tool and methodology for the analysis and control of solar cell structures and interfaces on the nano-scale.
As a result a new complete process chain, which potentially can substitute the conventional Si wafer based solar cell processing technology has been developed and disseminated. It is important that the results of the project, and the new capabilities of the partners, have been disseminated at conferences and exhibitions and were a part of the partners’ own strategies. Separate ThinSi dedicated work-shop "Advanced concepts for Si based PV", which was organised in collaboration with EU project nanoPV, R2M-Si and Hipersol has been an important tool in this respect.
Important to note that none of the individual partners or participating countries would be able to succeed in the development of the proposed technology new process flow based on a “powder-to-substrates” concept.
Dissemination activities and exploitation
It was supposed that the interest of the industrial partners and European industry in general will be protected by securing the IPR through patents. Since project time has been limited and all partners were overloaded by on-going developments, no patent applications were filed so far. Nevertheless, several patent applications are under preparation currently (thermal spray related). Patent applications will be submitted in the nearly future. Such situation prevented, partially, dissemination of some important results, since IPR have to be secured. After securing the IPR, the project partners will disseminate much more knowledge outside the Consortium, to allow other actors to contribute to technology development and deployment.
Confidential information provided by any of the partners was used freely within the scope and for the duration of the project. All information provided to the project team (or part thereof) was assumed to be 'Commercial in Confidence' and the provider loosed no rights through its disclosure to the project team.
To assure successful exploitation of the ThinSi scientific and technical results, dissemination and exploitation activities was running in parallel with the coordination of the project so that all activities were linked to successive knowledge transfer. Disseminating and exploiting ThinSi achievements were considered as an integral part of all workpackages and a responsibility for all partners. Integrity and consistency of all dissemination and exploitation efforts was supervised in WP5 - Dissemination and exploitation. The efforts allocated to these activities were take place along the following main lines:
1. Dissemination of results obtained: All partners were collaborating in the dissemination of the results among the research community, the photovoltaics community and their respective industries and markets in a broad sense, including other FP7 programme participants.
2. Exploitation: Industrial take-up of the technology developed in ThinSi played a central part in the project. To maximise the likelihood of success, relevant project partners prepared an “exploitation assistance package”. The package contains manufacturing procedures, infrastructure and IPR.
The planning of dissemination activities, which is a horizontal activity along the overall project lifecycle, was started immediately in the first project meeting. Dissemination policies were based on three major dissemination channels. Each dissemination activity was designed as a blend of dissemination activities from one or more channels, with respect to the target group(s). The three channels and their components (in bold) are:
a. Online dissemination: A project website provided a first access point for interested scientific and business parties into the ThinSi project. The objective of the website is to create a community of interested parties around the project, to accelerate their involvement and to create awareness of the results.
b. Non-electronic dissemination: Classical vehicles of knowledge transfer such as open project seminars intended for industrial end-users, articles in technical journals, peer-reviewed publications in scientific journals and presentations in conferences were focused on the dissemination of project results, mainly to experts and professionals.
c. Personal dissemination: As described in ThinSi DoW, each partner represented a network of contacts, technology and market overview that is much greater than the consortium itself. The scientists and engineers working for ThinSi were actively seeking to disseminate the results through contacts in formal networks (e.g. participation in other projects and involvement in organisations) and in their personal informal networks.
Plans for use of results
The ThinSi partners exploited the results in many directions. The research partners (SIN, ISE, IMEC and ENEA) were utilizing the academic potential in scientific publications and as basis for new research. They will also offer product and process development to industries. Two industrial partners, OIPT and IMPT, offered highly sophisticated production equipment (vacuum and non-vacuum techniques, respectively) to the global PV fabrication market. Another industrial partner, NTMD, offered sophisticated analysis equipment. ELKS offered specialised Si powder qualities for the new PV substrate technology. PYRG is going commercialise technology based on thermal spray of Si powder for processing of solar cells.
Exploitation for each partner
SINTEF: With the results from ThinSi, SINTEF is able to offer a comprehensive package of knowledge about low-cost Si substrates for solar cells to industrial companies on a world wide scale. Through industrial development projects, SINTEF offers assistance to the companies in setting up their own production facilities. SINTEF is also able to produce small series of substrates in our own laboratories. To develop SINTEF's capabilities further, SINTEF intends to continue the cooperation with the ThinSi partners in new research projects inside and outside the European framework programmes. In particular, main developments, which were realised in frame of ThinSi is transferred to new EU project application – Waste2PV (on a second stage evaluation at the EC) in which So powder based concept is a central line of the project activity. Pyrogenesis and IMEC are the partners in this new application coordinated by CEA (France). Moreover, Si powder-to substrate concept is a part of the development in a new EU project EuruSunMed, coordinated by CNRS (SINTEF is a partner). As a marketing tool for services and research capabilities SINTEF intends to publish the project’s results – after securing IPR – in high quality peer-reviewed journals, at relevant conferences, and in application-focused technical journals.
ISE Freiburg: High throughput silicon deposition and epitaxy is a key process for all c-Si thin-film solar cell approaches. Within ThinSi, ISE has optimized the ConCVD processes/tools, and adapted silicon epitaxy to a cell concept with best cost perspectives. ISE is planning to exploit these results by licensing the technology to industrial partners, who will be able to secure a quick ramp-up of a cell production. Small-series production of c-Si thin-film solar cells/wafer equivalents are possible also at ISE in its new SIMTEC lab. The equipment and processes will be used in accompanying and further R&D projects, also in continued projects with the ThinSi partners.
The work on advanced solar cell processes like SiC hetero emitters, epitaxial emitters or plasma texture not only has applications for the ThinSi approaches, but can be used for other silicon solar cell approaches as well. Depending on the status of each process, exploitation is planned in connecting public R&D projects and/or industrial projects. As usual for the Fraunhofer philosophy, close cooperation with industrial partners is planned to disseminate as much as possible of the know-how for industrial application.
In addition to the exploitation issues, ISE has disseminated the open/secured knowledge in publications and workshops, using the whole variety of publication channels.
IMEC: When the project is successfully finished, IMEC will be able to offer a comprehensive knowledge portfolio on thin film solar cell manufacturing on low-cost substrates to industrial companies on a world wide scale. The technique developed in this project nicely complements our existing expertise on thin film epitaxial Si cells on low-cost multi-crystalline substrates, and our knowledge on very thin film poly-crystalline cells on glass and ceramic substrates.
As a marketing tool for our services and research capabilities IMEC, likewise SINTEF intends to publish the project’s results – after securing IPR – in high quality peer-reviewed journals, at relevant conferences, and in application-focused technical journals.
To develop our capabilities further, we intend to continue the cooperation with the ThinSi partners in new research projects inside and outside the European framework programmes.
ENEA: Deeper understanding of growth specifics of layer for solar cell applications through optical characterization has been achieved. Methodology for optical test on single films and multilayer stacks has been elaborated with the stress on the complexity of mathematical modelling for such devices. The elaborated approaches are thought to be useful both for further exploitation in solar cell technology development and for similar applications (TCO-including multilayers). Hence further cooperation with the project partners is auspicious and will be fruitful thank to the basis for systematic and reliable characterization thought to be created within this project, in terms of both methodology and relevant optical benches construction. The scientific results of this research have been presented at high-level international conferences and in peer-reviewed journals.
University of Nottingham: The science and technology generated from the study of the new formulation and deposition of low-cost TCOs based on doped ZnO in ThinSi project allowed for UNOTT to secure new IPs in this field and transferred the knowledge and technology to SME such as IMPT for scaling up the ESAVD process for large area non vacuum deposition of low cost TCOs onto 8” Si wafers for solar cell applications.
In addition, UNOTT has been committed to the dissemination of non IP sensitive information to both academic and industry in UK and Europe via seminar, workshop, international conferences, annual report, and reputable peer reviewed journals. UNOTT also intends to continue to work with ThinSi partners in other new research projects inside and outside the European framework programmes.
Oxford Instruments Plasma Technology Ltd: OIPT can offer qualified tools and processes for at least two process steps in the fabrication of thin film PV devices. It has an established position as an international equipment supplier to both the R&D sector and to specific production niches, and is capable of growing with the opportunity this project presents. Its minimum expectation is to increase company turnover by 10% (from approximately 40M euro) by equipping a few PV manufacturers. There is clearly scope for this project to significantly impact the future of the business.
Elkem Solar AS: Elkem Solar produces Solar grade silicon through a multi-step metallurgical process route. In this route silicon fines side-streams of different quality are available. The primary goal of Elkem Solars participation in the project is to increase its knowledge of these side-streams and their potential as substrates towards thin film solar cells. Furthermore, Elkem Solar wishes expend its network in the area of thin film solar cells.
PyroGenesis S.A.:ThinSi project is of great importance for PyroGenesis from the exploitation of results point of view. It is an opportunity for Pyrogenesis to explore the application of thermal spray coatings in a new and rapidly developing sector, the sector of Solar Energy Materials. Thermal spray technologies are not applied for the moment for such applications on a commercial basis. Therefore, PyroGenesis is the first company worldwide, who has develop know-how on thermal spray coatings for PV applications. Exploitation plans for PyroGenesis include the commercialization of the successful coatings as results of the project. PyroGenesis will have a serious role in the co-production of the final products, together with the other partners of the project. The company wishes to participate in the commercial production, as far as thermal spray is concerned. PyroGenesis is also willing to examine the feasibility of the installation of a full production chain in Greece, either alone or in collaboration with the project’s partners or with new partners who could obtain rights on the final results.
NT-MDT Europe BV: NT-MDT will bring to the European market two unique tools for monitoring of electrical and structural properties of individual solar cell materials and interfaces. The first one is an analytical instrument based on Atomic Force Microscopy (AFM) and Spectral Optical methods and instruments for solar cell processing on nanoscale level. The second system is a combination of AFM and ellipsometry for the investigation of the physical characteristics thin films and interfaces of the collar elements.
IMPT Ltd.: After a successful implementation of ThinSi it is envisaged that the commercialisation through the sales of specialised equipment and/or licensing of the ESAVD process for the non-vacuum deposition of TCOs to solar cell industry will follow. This will open up significant opportunities and generate enormous benefits, and revenues to IMPT.
IMPT adopted the formulation and deposition of low-cost TCOs (e.g. FTO) developed by UNOTT and applied them onto its specialised scaled up ESAVD, and demonstrate the technical and commercial viability of the manufacture of such low-cost new TCOs by the non-vacuum ESAVD deposition onto large area substrates.
The ThinSi project provided a platform for IMPT to continue to collaborate with other partners to develop and apply the non vacuum ESAVD deposition method for the cost-effective fabrication of high quality thin and coatings for other engineering and device applications, within and outside the European framework programmes.
Isofoton S.A.: Isofoton considers the use of low cost silicon substrates as a requirement for the future of the photovoltaic energy. Within the ThinSi project, Isofoton accumulated an important knowledge about the results of the introduction of low purity substrates into laboratory scale processing lines. On the basis of such results and cost estimations, an evaluation of a new low-cost technology has been done at Isofoton.
List of Websites:
4.1.5 Project data
The ThinSi project has an open website available on http://www.sintef.no/Projectweb/ThinSi(si apre in una nuova finestra). On this website information about the project vision, goals and events are published. The coordinating organisation was SINTEF, and the coordinator was Alexander Ulyashin (alexander.g.ulyashin@sintef.no, +47 93002224). The logo of the project is shown below:
ThinSi Consortium
Beneficiary Number * Beneficiary name Beneficiary short name Country Contact person
1 (Coordinator) Stiftelsen SINTEF SIN Norway Alexander Ulyashin
alexander.g.ulyashin@sintef.no
2 Fraunhofer Institute for Solar Energy Systems ISE Germany Stefan Reber
stefan.reber@ise.fraunhofer.no
3 Interuniversitair Micro-Electronica centrum VZW IMEC Belgium Ivan Gordon
gordoni@imec.be
4 Ente per le Nuove tecnologie, l’Energia e l’Ambiente ENEA Italy Anna Sytchkova
annak@enea.it
5 University of Nottingham UNOTT UK Kwang-Leong Choy
Kwang-Leong.Choy@nottingham.ac.uk
6 Oxford Instruments Plasma Technology Ltd. OIPT UK Mike Cooke
Mike.COOKE@oxinst.com
7 Elkem Solar AS ELKS Norway Ronny Gløckner
Ronny.Gloeckner@elkem.no
8 PyroGenesis S.A. PYRG Greece Michalis Vardavoulias
mvardavoulias@pyrogenesis-sa.gr
9 NT-MDT Europe BV NTMD Netherlands Julia Alexeeva
ualexeeva@ntmdt.eu
10 Innovative Materials processing Technologies Ltd. IMPT UK Admin/finance:
Corinna Pinfold
Corinna.pinfold@imptl.com
Technical:
Keith Wells
keith@imptl.com
11 Isofoton S.A. ISOF Spain Miguel Vazquez
m.vazquez@isofoton.com
